Graded unmanned aerial vehicle take-off and landing site selection method for urban low-altitude logistics demand

Abstract

The invention discloses a graded unmanned aerial vehicle take-off and landing site selection method for urban low-altitude logistics demands, relates to infrastructure site selection, and aims to improve the distribution efficiency of urban logistics unmanned aerial vehicles and promote low-altitude economic development. Firstly, urban low-altitude logistics demands are evaluated, and key factors influencing the low-altitude logistics demands are captured; secondly, determining unmanned aerial vehicle take-off and landing point candidate points by applying a multi-stage facility polygon intersection point set method; thirdly, establishing a multi-target grading unmanned aerial vehicle take-off and landing point maximum coverage site selection model by maximizing the demand coverage rate, minimizing the construction cost and the demand repeated coverage rate; and finally, a QGPS-NSGA-III algorithm with improved design is adopted to solve the site selection model. According to the method, key factors influencing site selection can be accurately identified, multiple optimization targets are met to the maximum extent, the rationality and feasibility of unmanned aerial vehicle take-off and landing site selection are remarkably improved, and scientific decision support is provided for planning and implementation of an urban low-altitude logistics system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of infrastructure site selection, and in particular to a hierarchical unmanned aerial vehicle landing site selection method for urban low-altitude logistics demand. BACKGROUND

[0002] With the huge development opportunity of low-altitude economy, it is estimated that the market size of low-altitude economy in China will reach 100 billion yuan by 2030. As an important part of this field, low-altitude logistics has obvious advantages in urban distribution, including breaking through environmental restrictions, improving distribution efficiency, and reducing labor costs. To establish a normalized and large-scale unmanned aerial vehicle logistics network, not only advanced unmanned aerial vehicle technology is needed, but also market demand needs to be clarified, and supporting facilities, especially the "landing site" of logistics distribution transfer, need to be improved. Therefore, the hierarchical unmanned aerial vehicle landing site selection for urban low-altitude logistics demand helps to form a multi-level and multi-node low-altitude logistics network, improve the application efficiency of unmanned aerial vehicles in urban logistics, and promote the development of low-altitude economy.

[0003] Looking at the existing research results, there are the following problems: First, the logistics demand evaluation and unmanned aerial vehicle landing site selection are mainly based on traditional overall logistics demand prediction, lacking independent evaluation of urban low-altitude unmanned aerial vehicle logistics demand, limiting the consideration of low-altitude logistics characteristics, and affecting the optimization of landing site layout and operation efficiency. Second, although some researches discuss the hierarchical site selection of unmanned aerial vehicle facilities, they do not consider the service range redundancy between facilities at different levels, which may lead to resource waste. Finally, the existing research does not consider the impact of regional difference factors such as economic development level and traffic accessibility on landing site selection. Therefore, the present application evaluates urban low-altitude logistics demand, divides unmanned aerial vehicle landing sites into two levels according to construction cost and coverage, maximizes demand coverage, minimizes construction cost and demand overlap, and considers the impact of regional differences to finally obtain an unmanned aerial vehicle landing site selection scheme. SUMMARY

[0004] The present application discloses a hierarchical unmanned aerial vehicle landing site selection method for urban low-altitude logistics demand, which aims to consider the impact of regional low-altitude logistics demand differences on unmanned aerial vehicle landing site selection, maximize landing site demand coverage, and minimize construction cost and demand overlap rate.

[0005] To achieve the above-mentioned purpose, the technical scheme provided by the present application is:

[0006] A hierarchical unmanned aerial vehicle landing site selection method for urban low-altitude logistics demand, comprising the following steps:

[0007] Step 1: Evaluate urban low-altitude logistics demand and capture key factors affecting low-altitude logistics demand;

[0008] Step 2: analyze the logistics demand distribution by applying the multi-level facility polygon intersection point set method to obtain preliminary candidate points, and eliminate unselectable candidate points to obtain the candidate point set of the UAV take-off and landing point;

[0009] Step 3: establish a multi-objective hierarchical UAV take-off and landing point maximum coverage location selection model by maximizing demand coverage, minimizing construction cost and demand repeated coverage;

[0010] Step 4: design an improved QGPS-NSGA-III algorithm to solve the location selection model, and obtain the hierarchical UAV take-off and landing point location selection result for urban low-altitude logistics demand.

[0011] To optimize the above technical solutions, the specific measures / limitations taken also include:

[0012] In step 1, first calculate the utility function of each city logistics distribution mode. Since customers choose logistics distribution according to utility theory, the utility U uv of customer individual u choosing distribution mode v is calculated as follows:

[0013] U uv =x uv β u

[0014] β u is the coefficient vector of customer individual u; x uv is the influence factor vector set, representing the preference of customer individual u choosing distribution mode v.

[0015] Secondly, the conditional Logit model is derived by assuming the generalized extreme value distribution form of the error term, and its standard form is as follows:

[0016]

[0017] In the formula, H uv is the standard conditional probability of customer individual u choosing distribution mode v.

[0018] Finally, the UAV distribution mode share rate is calculated, and the UAV distribution mode share rate of all logistics distribution records is spatially accumulated to obtain the spatial distribution of urban low-altitude logistics demand.

[0019] In step 2, the multi-level facility polygon intersection point set method is applied to analyze the logistics demand distribution, and the unselectable candidate points are eliminated to obtain the candidate point set of the UAV take-off and landing point, and the steps are as follows:

[0020] Step 1: define the facility candidate area under PIPS method when single-level facility location selection is performed, obtain the facility candidate area of demand polygon, and obtain the PIPS candidate point through the intersection relationship of facility candidate areas of different demand polygons;

[0021] Step2: The PIPS candidate points and the dominance relationship are extended to hierarchical facility location, and the UAV landing points are divided into two levels of landing points and landing fields. The city low-altitude logistics demand distribution is further analyzed to obtain the HFPIPS candidate points.

[0022] Step3: The candidate points located in the city unselectable area are removed, and the final set of UAV landing point candidate points is obtained.

[0023] Considering the uneven distribution and large volume difference of city low-altitude logistics demand, the UAV landing points are divided into two levels according to service level, construction cost and coverage:

[0024] UAV landing point (level 1): service coverage radius 5km; construction cost 50000 yuan; capacity 20000;

[0025] UAV landing field (level 2): service coverage radius 8km; construction cost 80000 yuan; capacity 50000;

[0026] In step 3, a multi-objective hierarchical UAV landing point maximum coverage location model is established.

[0027] Considering maximizing the coverage rate of city low-altitude logistics demand, the objective function formula is as follows:

[0028]

[0029] Where I is the set of city low-altitude logistics demand points, w i is the demand of low-altitude logistics demand point i, s i is the decision variable, which represents whether the demand point is covered by the facility, if covered by the facility s i =1, otherwise s i =0;

[0030] Considering minimizing the construction cost of UAV landing point, the objective function formula is as follows:

[0031]

[0032] J is the set of UAV landing point candidate points, c1 is the construction cost of landing field, c2 is the construction cost of landing point, x j , y j is the decision variable, which represents whether to build a landing point (landing field) at candidate point j, if 1, otherwise 0;

[0033] Considering minimizing the city low-altitude logistics demand repeated coverage rate, the objective function formula is as follows:

[0034]

[0035] z i is a decision variable, representing whether demand point i is covered by a facility repeatedly, z i = 1 if demand point i is covered by a facility repeatedly, otherwise z i = 0.

[0036] To adapt to the hierarchical location requirements, the constraints cover multiple aspects such as coverage constraints, number of take-off and landing site constraints, capacity constraints, service range constraints, etc., to ensure that the model is as realistic as possible. At the same time, in order to optimize resource allocation, meet the demand of low-altitude logistics, and reduce the waste of construction resources, the model also makes many assumptions to simplify the difficulty of the model, which are as follows:

[0037] (1) Low-altitude logistics distribution is completed by vertical take-off and landing rechargeable rotor unmanned aerial vehicles, and all unmanned aerial vehicles fly at a constant speed.

[0038] (2) The impact of goods types on distribution is not considered, and all order distribution has the same priority.

[0039] (3) The unmanned aerial vehicle take-off and landing point can serve multiple demand objects.

[0040] (4) The take-off and landing point meets the unmanned aerial vehicle distribution requirements.

[0041] In step 4, the improved QGPS-NSGA-III algorithm is designed to solve the location model:

[0042] First, generate an initial solution that meets the spacing constraints between take-off and landing points based on the grid special point set method. Define a grid with a side length of 3 km, and divide the HFPIPS points by location; To ensure that the initial solution meets the model constraints and improve the fitness, calculate the average low-altitude logistics demand coverage rate of the HFPIPS points in each grid, and sort them according to this rate. Select the top 1500 grids to form the "high-quality grid set"; Randomly select a grid from the set and generate a 0-1 random number. If it is less than 0.5, sort the candidate points by the demand coverage rate of the take-off and landing point, otherwise sort them by the demand coverage rate of the take-off and landing site; Select the HFPIPS point with the highest demand coverage rate and no airspace conflict to add to the initial solution. If there is no suitable point, skip it; Repeat this process until the desired number of initial solutions is reached.

[0043] Second, construct uniformly distributed reference points on the standard hyperplane. Construct uniformly distributed reference points on the M-1 dimensional standard hyperplane, which intersects the coordinate axis and has an intercept of 1; The number of reference points H is related to the target dimension and population size. To maintain population diversity, the number of reference points should be equal to or slightly less than the population size.

[0044] Then, the excellent individuals are selected from the population to form a mating pool, and single-point crossover and multi-point mutation are used for crossover and mutation operations. A random number of 0-1 is randomly generated, if greater than 0.5, a mutation is selected from all HFPIPS candidate point sets to improve the global optimization ability of the algorithm and avoid falling into local optimum, if less than 0.5, a mutation is selected from the high-quality grid point set to improve the local optimization ability of the algorithm, after the crossover and mutation operations, it is checked whether the offspring chromosome exists repeated coding or violates the constraint problem and is repaired.

[0045] Finally, adaptive population normalization is performed, the individual is associated with the reference point, and the elite reservation operation is performed.

[0046] Compared with the prior art, the beneficial effects of the present application are:

[0047] The present application aims at the deficiencies in the existing research on the site selection of the landing point of the urban logistics unmanned aerial vehicle, and a landing point site selection model of the unmanned aerial vehicle is constructed by comprehensively considering demand coverage, construction cost and demand repeated coverage, and the QGPS-NSGA-III algorithm is improved and designed to solve the hierarchical landing point site selection problem of the unmanned aerial vehicle for urban low-altitude logistics demand. Compared with other models, the conditional Logit model is used to identify the key factors affecting the logistics demand, and the effective separation of the urban low-altitude logistics demand is realized. At the same time, the model and the algorithm can be applied to the site selection of the landing point of the urban logistics unmanned aerial vehicle, and the spatial distribution of the urban low-altitude logistics demand is effectively targeted, the hierarchical landing point of the unmanned aerial vehicle is reasonably set, and the resource allocation is optimized. It can not only ensure a high service coverage rate, but also reduce the construction cost and reduce the redundancy of the service range. The improved QGPS-NSGA-III algorithm is superior to the NSGA-II algorithm in solving the maximum coverage site selection model of the multi-objective hierarchical landing point of the unmanned aerial vehicle, and especially shows significant advantages in demand coverage rate and construction cost. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 is a flow chart of the method of the present application.

[0049] Figure 2 is a low-altitude logistics demand distribution map of a certain city.

[0050] Figure 3 is a distribution map of the candidate point of the landing point of the unmanned aerial vehicle of a certain city.

[0051] Figure 4 is a demand coverage result map of the case unmanned aerial vehicle landing site. DETAILED DESCRIPTION

[0052] The above content of the present application is further described in detail by way of examples, but this should not be understood as limiting the scope of the above subject matter of the present application to the following examples only, and any technology realized based on the above content of the present application falls within the scope of the present application.

[0053] The present application proposes a hierarchical unmanned aerial vehicle landing site selection method for urban low-altitude logistics demand, and a flow chart is shown as Figure 1 The method comprises the following steps:

[0054] (1) Considering the distribution of urban low-altitude logistics demand, the key factors affecting low-altitude logistics demand are captured.

[0055] The specific steps in step (1) include:

[0056] 1.1, calculate the utility function of each city logistics distribution mode. Since customers choose logistics distribution according to utility theory, the utility U uv of customer individual u choosing distribution mode v is calculated as follows:

[0057] U uv =x uv β u

[0058] β u is the coefficient vector of customer individual u; x uv is the impact factor vector set, representing the preference of customer individual u choosing distribution mode v.

[0059] The conditional Logit model is derived by assuming a generalized extreme value distribution form of the error term, and its standard form is as follows:

[0060]

[0061] In the formula, H uv is the standard conditional probability of customer individual u choosing distribution mode v.

[0062] 1.2, calculate the sharing rate of unmanned aerial vehicle distribution mode, and spatially accumulate the sharing rate of unmanned aerial vehicle distribution mode for all logistics distribution records to obtain the spatial distribution of urban low-altitude logistics demand.

[0063] (2) Apply the multi-level facility polygon intersection point set method to analyze the logistics demand distribution, and remove the unselectable candidate points to obtain the unmanned aerial vehicle landing site candidate point set.

[0064] The specific steps in step (2) include:

[0065] 2.1, define the PIPS method under the facility candidate area of single-level facility location, get the demand polygon facility candidate area, and get the PIPS candidate point through the intersection relationship of different demand polygon facility candidate areas.

[0066] 2.2, extend the PIPS candidate point and dominance relationship to hierarchical facility location, divide the UAV landing point into two levels of landing point and landing field, further analyze the distribution of urban low-altitude logistics demand, and get the HFPIPS candidate point.

[0067] 2.3, eliminate the candidate points located in the urban non-selectable area, and finally get the UAV landing point candidate point set.

[0068] Considering the uneven distribution and large volume difference of urban low-altitude logistics demand, the UAV landing point is divided into two levels according to service level, construction cost and coverage range:

[0069] UAV landing point (level 1): service coverage radius 5km; construction cost 50000 yuan; capacity 20000;

[0070] UAV landing field (level 2): service coverage radius 8km; construction cost 80000 yuan; capacity 50000;

[0071] (3) A multi-objective hierarchical UAV landing point maximum coverage location model is established to maximize demand coverage rate and minimize construction cost and demand repeated coverage rate.

[0072] The step (3) describes the model:

[0073] Considering the maximization of urban low-altitude logistics demand coverage rate, the objective function formula is as follows:

[0074]

[0075] Where I is the set of urban low-altitude logistics demand points, w i is the demand of low-altitude logistics demand point i, s i is the decision variable, which represents whether the demand point is covered by the facility, if it is covered by the facility s i =1, otherwise s i =0;

[0076] Considering the minimization of UAV landing point construction cost, the objective function formula is as follows:

[0077]

[0078] J is the set of UAV landing point candidate points, c1 is the construction cost of landing field, c2 is the construction cost of landing point, x j , y jis a decision variable, representing whether to build a take-off and landing point (airport) at candidate point j. If a take-off and landing point is built at candidate point j, it is 1, otherwise it is 0;

[0079] Considering the minimization of the repeated coverage rate of urban low-altitude logistics demand, the objective function formula is as follows:

[0080]

[0081] z i is a decision variable, representing whether demand point i is repeatedly covered by the facility service. If demand point i is repeatedly covered by the facility service, z i = 1, otherwise z i = 0.

[0082] To adapt to the requirements of hierarchical location selection, the constraint conditions cover multiple aspects such as coverage constraint, airport point number constraint, capacity constraint, service range constraint, etc., to ensure that the model is as real as possible. At the same time, in order to optimize resource allocation, meet the demand of low-altitude logistics, and reduce the waste of construction resources, the model also makes many assumptions to simplify the difficulty of the model, and the assumptions are as follows:

[0083] (a) The low-altitude logistics distribution work is completed by a vertical take-off and landing rechargeable rotor unmanned aerial vehicle, and all unmanned aerial vehicles fly at a uniform speed.

[0084] (b) The impact of cargo types on distribution is not considered, and all order distribution has the same priority.

[0085] (c) The unmanned aerial vehicle take-off and landing point can serve multiple demand objects.

[0086] (d) The unmanned aerial vehicle take-off and landing point meets the unmanned aerial vehicle distribution requirements.

[0087] (4) Design an improved QGPS-NSGA-III algorithm to solve the location selection model.

[0088] The main steps of the algorithm in step (4) include:

[0089] 4.1, generate an initial solution that meets the distance constraint between take-off and landing points based on the grid special point set method:

[0090] Define a grid with a side length of 3 km, and divide the HFPIPS points by position; To ensure that the initial solution meets the model constraints and improve the fitness, calculate the average low-altitude logistics demand coverage rate of the HFPIPS points in each grid, and sort them according to this rate. Select the top 1500 grids to form a "high-quality grid set"; Randomly select a grid from the set, and generate a 0-1 random number. If it is less than 0.5, sort the candidate points according to the demand coverage rate of the take-off and landing point, otherwise sort them according to the demand coverage rate of the take-off and landing field; Select the HFPIPS point with the highest demand coverage rate and no air space conflict to add to the initial solution. If there is no suitable point, skip it; Repeat this process until the desired number of initial solutions is reached.

[0091] 4.2, constructing evenly distributed reference points on the standardized hyperplane:

[0092] Construct evenly distributed reference points on the M-1 dimensional standardized hyperplane, which intersects the coordinate axis and has an intercept of 1; the number of reference points H is related to the target dimension and the population size, and to maintain population diversity, the number of reference points should be equal to or slightly less than the population size.

[0093] 4.3, selecting excellent individuals from the population to form a mating pool, and performing crossover and mutation operations using single-point crossover and multi-point mutation:

[0094] Randomly generate a random number between 0 and 1, if greater than 0.5, select mutation from all HFPIPS candidate point sets to improve the global optimization ability of the algorithm and avoid falling into local optimum; if less than 0.5, select mutation from the high-quality grid point set to improve the local optimization ability of the algorithm; after crossover and mutation operations, check whether the offspring chromosomes have repeated coding or violate the constraints and repair them.

[0095] 4.4, adaptive population normalization, link individuals with reference points, perform elite preservation operation, and get hierarchical unmanned aerial vehicle landing site selection results for urban low-altitude logistics demand.

[0096] The technical solutions of the present application will be further illustrated by a specific embodiment.

[0097] The logistics distribution data used in this case mainly comes from the monthly logistics distribution order data of a certain city provided by Cainiao Network, covering 124 network points and related package data. The unmanned aerial vehicle distribution system of the city consists of three layers: logistics distribution center, unmanned aerial vehicle landing site and landing point, and express station. The logistics distribution center is responsible for centralized processing and distribution of packages, and delivers packages to unmanned aerial vehicle landing sites and landing points through cargo transport vehicles. Unmanned aerial vehicle landing sites (large landing points) have high processing capacity and service level, and are suitable for densely populated areas, while unmanned aerial vehicle landing points are flexibly set up at specific locations to serve small-scale distribution needs. Express stations serve as the final destination and are responsible for receiving and distributing packages.

[0098] (1) Distribution analysis of low-altitude logistics demand in a certain city

[0099] The utility function parameters are solved using the conditional Logit model, and the results are shown in Table 1. The results show that an increase in delivery distance has a positive effect on urban low-altitude logistics demand, while an increase in delivery time and cost significantly reduces demand, indicating that customers are highly sensitive to time and cost. The Logit model performs well overall, effectively capturing key factors affecting low-altitude logistics demand. Next, the entire region of the city is divided into 2 km * 2 km square grids, resulting in a total of 1910 grids. The low-altitude logistics demand distribution of the city obtained by spatial aggregation is shown in Figure 2 .

[0100] Table 1 Solution results of conditional Logit model

[0101]

[0102] (2) Distribution analysis of UAV landing point candidate points in a certain city

[0103] The city's UAV landing point non-selectable areas, including water areas, residential areas, and public management service land, are determined based on the city's water system data obtained from Open StreetMap and the city's land use category data provided by EULUC-China. After analyzing the city's low-altitude logistics demand distribution using the HFPIPS method, the candidate points located in the non-selectable areas are excluded, and the final set of UAV landing point candidate points in the city is shown in Figure 3 .

[0104] (3) UAV landing point site selection results in a certain city

[0105] The evaluation results of the city's low-altitude logistics demand in the first step and the UAV landing point candidate point results in the second step are used as inputs to solve the multi-objective hierarchical UAV landing site selection model. Python is used to program the QGPS-NSGA-III algorithm, which is applied to the solution of the multi-objective hierarchical site selection model. Considering that the invention is aimed at the multi-objective UAV landing site selection problem, multi-objective decision-making is used to assist in reasonably selecting the optimal solution of the multi-objective hierarchical site selection scheme. Considering the initial construction characteristics of UAV landing points and the strong support of the state, according to the prediction of the UAV logistics distribution mode in previous research, the decision target weight scheme shown in Table 2 is designed. The specific idea of this decision is that during the initial construction of UAV landing points, a larger low-altitude logistics demand coverage and a smaller demand overlap are relatively important, while the construction cost accounts for a smaller proportion.

[0106] Table 2 Decision target weight scheme

[0107]

[0108] According to the results of sensitivity analysis, the solution results of QGPS-NSGA-III algorithm with the construction quantity P=160 are used as the site selection scheme of the unmanned aerial vehicle landing site for the urban low-altitude logistics demand, and the site selection results and demand coverage are shown in Figure 4 A total of 160 unmanned aerial vehicle landing sites are constructed, of which 53 are unmanned aerial vehicle landing sites and 107 are landing sites. Considering the initial capital limitation and resource waste, the scheme finally achieves a demand coverage rate of 84.5%, a construction cost of 9.59 million yuan, a demand repeated coverage rate of 33.6%, and has rationality and feasibility.

[0109] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Any skilled person in the art, without departing from the technical solution of the present application, according to the technical essence of the present application, any simple modification, equivalent replacement and improvement of the above embodiment, etc. still belongs to the protection scope of the technical solution of the present application.

Claims

1. A hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs, characterized by: The following steps are involved: Step 1: Assess the demand for low-altitude logistics in cities and capture the key factors that affect low-altitude logistics demand; Step 2: Apply the multi-level facility polygon intersection point set method to analyze the logistics demand distribution to obtain preliminary candidate points, and eliminate unselectable candidate points to obtain the candidate point set for drone take-off and landing points; Step 3: Establish a multi-objective hierarchical UAV take-off and landing point maximum coverage site selection model to maximize demand coverage, minimize construction cost and demand duplication coverage; Step 4: Design an improved QGPS-NSGA-III algorithm to solve the site selection model and obtain the hierarchical drone take-off and landing point site selection results for urban low-altitude logistics needs.

2. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 1 is characterized in that: The specific process of step 1 is as follows: Step 1.1: Calculate the utility function of each city's logistics distribution model; Since customers follow the utility theory to choose logistics distribution, the utility U of individual customer u when choosing distribution mode v is uv The calculation formula is as follows: U uv =x uv b u β u is the coefficient vector of individual u; x uv is a set of influencing factor vectors, representing the preference of individual customer u for choosing delivery mode v; The conditional logit model is derived by assuming the generalized extreme value distribution of the error term. Its standard form is as follows: Where H uv The standard conditional probability of customer individual u choosing delivery mode v; Step 1.2: Calculate the share of the drone delivery mode and spatially accumulate the share of the drone delivery mode of all logistics delivery records to obtain the spatial distribution of urban low-altitude logistics demand.

3. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 1 is characterized in that: The specific method of step 2 is as follows: Step 2.1: Define the facility candidate area under the PIPS method for single-level facility site selection, obtain the facility candidate area of ​​the demand polygon, and then obtain the PIPS candidate point through the intersection relationship of the facility candidate areas of different demand polygons; Step 2.2: Extend the PIPS candidate points and dominance relationships to the hierarchical facility site selection, divide the drone take-off and landing points into two levels: take-off and landing points and take-off and landing fields, further analyze the distribution of urban low-altitude logistics demand, and obtain HFPIPS candidate points; Step 2.3: Eliminate candidate points located in unavailable urban areas, and finally obtain the candidate point set for drone take-off and landing points.

4. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 3 is characterized in that: In step 2.2, considering the uneven distribution of urban low-altitude logistics demand and large differences in volume, drone take-off and landing points are divided into two levels based on service level, construction cost, and coverage: Level 1 drone take-off and landing point: service coverage radius 5km; construction cost 50,000 yuan; capacity 20,000; Level 2 drone landing and take-off site: service coverage radius 8km; construction cost 80,000 yuan; capacity 50,000.

5. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 1 is characterized in that: In step 3, the multi-objective hierarchical UAV take-off and landing point maximum coverage site selection model is as follows: Considering maximizing the coverage rate of urban low-altitude logistics demand, the objective function formula is as follows: Among them, I is the set of urban low-altitude logistics demand points, w i is the demand for low-altitude logistics demand point i, s i is a decision variable, representing whether the demand point is covered by the facility service. If it is covered by the facility service, s i =1, otherwise s i =0; Considering minimizing the construction cost of drone take-off and landing points, the objective function formula is as follows: J is the set of candidate points for the take-off and landing points of the UAV, c1 is the construction cost of the take-off and landing field, c2 is the construction cost of the take-off and landing point, x j 、y j is a decision variable, representing whether to build a take-off and landing point at candidate point j. If so, it is 1, otherwise it is 0; Considering minimizing the repeated coverage rate of urban low-altitude logistics demand, the objective function formula is as follows: z i is a decision variable, representing whether demand point i is repeatedly covered by facility services. If it is repeatedly covered by facility services, then z i =1, otherwise z i =0; To meet the requirements of hierarchical site selection, the constraints include coverage constraints, take-off and landing site number constraints, capacity constraints, and service range constraints to ensure the authenticity of the model.

6. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 5 is characterized in that: In order to optimize resource allocation, meet low-altitude logistics needs, and reduce construction resource waste, the model also makes many assumptions to simplify the model difficulty. The assumptions are as follows: (1) Low-altitude logistics and distribution work is completed by rechargeable rotor drones that can take off and land vertically, and all drones fly at a constant speed; (2) Regardless of the impact of the type of goods on delivery, all orders have the same delivery priority; (3) The drone take-off and landing points can serve multiple demand objects; (4) All cargo at the take-off and landing points meet the drone delivery requirements.

7. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 1 is characterized in that: The specific process of the improved QGPS-NSGA-III algorithm in step 4 is as follows: Step 4.1: Generate an initial solution that satisfies the UAV take-off and landing point spacing constraints based on the grid special point set method and achieve effective mutation; Step 4.2: Construct uniformly distributed reference points on the normalized hyperplane; Step 4.3: Select excellent individuals from the population to form a mating pool, and use single-point crossover and multi-point mutation to perform crossover and mutation operations; Step 4.4: Perform adaptive population normalization, link individuals to reference points, and perform elite retention operations to obtain the hierarchical drone take-off and landing point selection results for urban low-altitude logistics needs.

8. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 7 is characterized in that: In step 4.1: Define a grid with a side length of 3 km and divide the HFPIPS points into locations. To ensure that the initial solution meets the model constraints and improves fitness, calculate the average low-altitude logistics demand coverage rate of HFPIPS points in each grid and sort them accordingly. Select the top 1500 grids to form a "high-quality grid set." Randomly select a grid from this set and generate a random number between 0 and 1. If the number is less than 0.5, the candidate points are sorted by the demand coverage rate of the take-off and landing point; otherwise, they are sorted by the demand coverage rate of the take-off and landing field. Select the HFPIPS point with the highest required coverage and no airspace conflict to add to the initial solution. If there is no suitable point, skip it. Repeat this process until the required number of initial solutions is reached.

9. The hierarchical drone take-off and landing point selection method for urban low-altitude logistics needs according to claim 7 is characterized in that: The specific process of crossover mutation in step 4.3 is as follows: When performing the mutation operation, a random number between 0 and 1 is randomly generated. If it is greater than 0.5, mutations are selected from all HFPIPS candidate points to improve the algorithm's global optimization ability and avoid falling into local optimality. If it is less than 0.5, mutations are selected from high-quality grid points to improve the algorithm's local optimization ability. After the crossover mutation operation, the offspring chromosomes are checked for duplicate encoding or constraint violations and repaired.

Images