Urban governance-oriented unmanned aerial vehicle nest site selection and layout method

Abstract

The invention relates to the technical field of unmanned aerial vehicle nest site selection, in particular to an unmanned aerial vehicle nest site selection and layout method for urban governance, and the method comprises the steps: dividing a target region, and obtaining the address of each to-be-screened nest; obtaining a comprehensive evaluation index according to the coverage condition of the population in the to-be-screened machine nest address, the concentration condition of the residence community, the concentration condition of various infrastructures and the distribution of residents interfered by the machine nests; according to the comprehensive evaluation index and distance distribution between the to-be-screened machine nest addresses, screening the to-be-screened machine nest addresses to obtain alternative addresses; obtaining a target function according to the coverage condition of the residence community in the unmanned aerial vehicle service range corresponding to each alternative address, the comprehensive evaluation index of the alternative address and the quantity distribution condition; and optimizing the target function by using a tabu search algorithm to obtain a nest site selection scheme and a corresponding number of unmanned aerial vehicle nests. The task execution efficiency and flexibility are improved, and the limitation of isolated arrangement in the traditional technology is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of drone nest site selection, and in particular to a method for drone nest site selection and layout for urban management. Background Art

[0002] In the context of smart cities, drones are increasingly used, especially in the field of urban governance. Drones can perform a variety of tasks including logistics distribution, urban management, and emergency response. In order to ensure that drones can complete these tasks efficiently and reliably, the reasonable location and layout of drone nests are particularly important. Drone nests are key nodes for drone takeoff and landing, charging, and data transmission. The rationality of their location and layout directly affects the coverage, response speed, and operating costs of drone services.

[0003] Existing technologies do not build drone nests into an interconnected network. Most current drone nests are independently and dispersedly arranged, lacking a systematic network layout plan. This isolated layout makes it difficult for drones to achieve efficient collaborative operations and resource sharing during mission execution, limiting the overall performance and expansion capabilities of the drone system, resulting in poor results in drone nest site selection. Summary of the invention

[0004] In order to solve the technical problem that the existing method for selecting the site of a drone nest is poor in effect, the purpose of the present invention is to provide a method for selecting and arranging the site of a drone nest for urban management. The technical solution adopted is as follows: Divide the target area to obtain the address of each machine nest to be screened, and the location of the machine nest to be screened meets the preset restriction conditions; According to the population coverage, residential community concentration, various infrastructure concentrations and the distribution of residents affected by the machine nest in the machine nest address to be screened, the machine nest address to be screened is evaluated to obtain a comprehensive evaluation index; According to the comprehensive evaluation index and the distance distribution between the addresses of the machine nests to be screened, the addresses of the machine nests to be screened are screened to obtain candidate addresses; The objective function is obtained based on the coverage of residential areas within the drone service range corresponding to each candidate address, the comprehensive evaluation index and quantity distribution of the candidate addresses; The taboo search algorithm is used to optimize the objective function to obtain the nest site selection plan and the corresponding number of UAV nests.

[0005] Preferably, the method for obtaining the address of the machine nest to be screened specifically includes: The target area is evenly divided into a dot grid with a preset length. A square area of ​​a preset size with each grid point on the dot grid as the center is the address of each drone nest. The drone nest address corresponding to the preset restriction condition is used as the nest address to be screened.

[0006] Preferably, the address of the machine nest to be screened is evaluated to obtain a comprehensive evaluation index, which specifically includes: Determine the first optimization index of each machine nest address to be screened based on the population coverage rate in the area where each machine nest address to be screened is located; Determine the second optimal index of each to-be-screened machine nest address based on the distribution density of residential communities in the area where each to-be-screened machine nest address is located; Determine the third optimization index of each to-be-screened machine nest address based on the POI kernel density of the area where each to-be-screened machine nest address is located; Determine the fourth optimization index of each machine nest address to be screened based on the NIMBY index of the area where each machine nest address to be screened is located; Based on the preset weight value, the first preferred index, the second preferred index, the third preferred index and the fourth preferred index are weightedly summed to determine the comprehensive evaluation index of each machine nest address to be screened, and the first preferred index, the second preferred index, the third preferred index and the fourth preferred index are all normalized values.

[0007] Preferably, the method of screening the machine nest addresses to be screened to obtain candidate addresses according to the comprehensive evaluation index and the distance distribution between the machine nest addresses to be screened specifically includes: The addresses of the machine nests to be screened corresponding to the comprehensive evaluation index being greater than the preset evaluation threshold are taken as candidate addresses, and the Euclidean distance between the particles in the areas where each two candidate addresses are located is within a preset distance range.

[0008] Preferably, the objective function is obtained according to the coverage of residential areas within the drone service range corresponding to each candidate address, the comprehensive evaluation index and quantity distribution of the candidate addresses, and specifically includes: Obtaining a reward item according to the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold; A first penalty item is determined based on the total number of all candidate addresses by weighting; a second penalty item is obtained based on the coverage of residential communities within the drone service range corresponding to each candidate address; and a comprehensive penalty item is determined based on the cumulative sum of the first penalty item and the second penalty item; The objective function is determined based on the difference between the reward term and the comprehensive penalty term.

[0009] Preferably, the reward item is obtained according to the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold, specifically including: The difference between the comprehensive evaluation index of each candidate address and the evaluation threshold is calculated, and the cumulative sum of the differences corresponding to all candidate addresses is weighted to obtain a reward item.

[0010] Preferably, the second penalty item is obtained according to the coverage of the residential area within the service range of the drone corresponding to each candidate address, and specifically includes: A circular area with the mass point of each candidate address as the center and the length of the drone working radius as the radius is used as the drone service range of each candidate address; Calculate the Euclidean distance between the mass point in the area where each residential community is located and the mass point of the alternative address. When the value of the Euclidean distance is greater than the radius of the drone service range, the corresponding residential community is not within the drone service range of the alternative address; when the value of the Euclidean distance is less than or equal to the radius of the drone service range, the corresponding residential community is within the drone service range of the alternative address. The second penalty term is obtained by weighting the total number of residential communities that are not within the service range of the drone at the alternative address.

[0011] The embodiments of the present invention have at least the following beneficial effects: The present invention covers multi-dimensional factors for the site selection of drones for urban governance, including population coverage, community core density, POI core density, NIMBY index, etc., forming a comprehensive and standard site selection evaluation system, which ensures that the site selection of drone nests is better and the evaluation is more accurate. Layout optimization is carried out from the macro perspective of the overall region or city, avoiding the shortcomings of traditional technology that is limited to local optimization. Through global optimization, it is possible to ensure the reasonable distribution of drone nests in the region, improve the efficiency and balance of the overall layout, and meet the most extensive service needs at the lowest economic cost. The drone nests are organized into an efficient interconnected network. This networked layout not only supports the flow of drones between different nests, but also improves the efficiency and flexibility of task execution, solves the limitations of isolated layout in traditional technology, and significantly enhances the overall effectiveness of the drone system. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0013] Figure 1It is a flowchart of the steps of a method for selecting and arranging drone nests for urban management provided by the present invention; Figure 2 It is a flowchart of the steps of the method for obtaining the comprehensive evaluation index of the machine nest address to be screened provided by the present invention; Figure 3 It is a flow chart of the steps of the method for constructing the objective function provided by the present invention; Figure 4 is a flowchart of the steps of optimizing the objective function provided by the present invention; Figure 5 It is a flow chart of constructing an initial solution provided by the present invention. DETAILED DESCRIPTION

[0014] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the method for selecting and arranging the site of drone nests for urban management proposed by the present invention, its specific implementation method, structure, features and effects, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0015] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0016] The following is a detailed description of a method for selecting and laying out drone nests for urban management provided by the present invention in conjunction with the accompanying drawings.

[0017] See also Figure 1 , which shows a flowchart of a method for selecting and arranging drone nests for urban management provided by an embodiment of the present invention, the method comprising the following steps: Step S100, dividing the target area to obtain the address of each machine nest to be screened, and the site selection of the machine nest to be screened meets the preset restriction condition.

[0018] Specifically, the target area is evenly divided into a dot grid with a preset length, and a square area of ​​a preset size with each grid point on the dot grid as the center is the address of each drone nest; the drone nest address corresponding to the preset restriction condition is used as the nest address to be screened.

[0019] The preset length is 100 meters and the preset size is , m represents meters. Each grid point in the dot grid has a corresponding longitude and latitude. For example, the longitude and latitude coordinates of the nth point can be expressed as .

[0020] Then, the restriction index is used to constrain the candidate addresses. The restriction index data is imported into the Geographic Information System (GIS). If it is raster data, it must be converted into vector data first and projected to the same geographic coordinate system as the grid points. Then, the restriction index is used to constrain the grid points. The restriction index includes no-fly zones, wind farms, and various electromagnetic wave transmitting base stations.

[0021] More specifically, the restrictions include: the grid point is more than 200 meters away from the no-fly zone, the average wind speed is less than 12m / s, and it is more than 200 meters away from the electromagnetic interference area, where the electromagnetic interference area includes radar stations, microwave stations, mobile communication base stations, drone interference equipment, etc. The area corresponding to the grid point that meets the restrictions is used as the machine nest address to be screened.

[0022] Step S200, evaluating the machine nest address to be screened according to the population coverage, the concentration of residential communities, the concentration of various infrastructures and the distribution of residents affected by the machine nest in the machine nest address to obtain a comprehensive evaluation index.

[0023] The embodiment of the present invention analyzes the characteristic distribution of data of each machine nest address to be screened in each dimension respectively, combines the data analysis results of multiple dimensions, and evaluates each machine nest address to be screened, so as to comprehensively consider multiple factors and quantitatively consider the optimal situation of the current machine nest site selection to be screened.

[0024] In this embodiment, the characteristic distribution of the address of the machine nest to be screened in four aspects is considered, and then the comprehensive evaluation result is determined, such as Figure 2 As shown, the method for obtaining the comprehensive evaluation index of each machine nest address to be screened can be implemented by steps S201 to S205.

[0025] Step S201, determining a first preferred index for each to-be-screened machine nest address based on the population coverage rate in the area where each to-be-screened machine nest address is located.

[0026] Specifically, the population coverage rate refers to the proportion of the number of people covered by the drone service at the target address to the total population. Taking any machine nest address to be screened as an example, the drone service range can be a circular area with the particle point of the machine nest address to be screened as the center and the length of the drone's working radius as the radius.

[0027] Furthermore, the calculation formula for the population coverage rate corresponding to the machine nest address to be screened can be expressed as:

[0028] in, represents the population coverage rate, Indicates the total population in the area where the machine nest address to be screened is located. Indicates the number of people within the service range of the drone of the machine nest address to be screened.

[0029] The population coverage rate corresponding to the machine nest address to be screened is normalized to obtain the first preferred address of the machine nest address to be screened, wherein the normalization method can be processed using the maximum and minimum value normalization method, which is a well-known technology and will not be introduced in detail here.

[0030] Step S202: determining a second preferred index for each to-be-screened machine nest address based on the distribution density of residential communities in the area where each to-be-screened machine nest address is located.

[0031] Construct a set of mass points for residential areas. For residential areas with a certain area, calculate their geometric center (mass point) or use the minimum enclosing rectangle method to idealize them into a single mass point, which represents the geographical location of the residential area on the map.

[0032]

[0033] Among them, A represents the mass point set consisting of n mass points in residential areas, Represents the kth residential area mass point. represents the longitude and latitude coordinates of the kth residential area. The coordinate matrix of the residential area mass point can be expressed as .

[0034] Specifically, the second preferred index can be obtained by calculating the core density of residential communities in the area where the address of the machine nest to be screened is located, that is, the core density of residential communities refers to the concentration degree of residential communities.

[0035] In this embodiment, the kernel density of the residential area corresponding to the address of the machine nest to be screened is calculated by using the Gaussian kernel function, which can be expressed as follows:

[0036] in, represents the kernel density of residential areas, is the bandwidth parameter, n represents the number of residential communities, represents the longitude and latitude coordinates of the mass point in the ith residential area, Indicates the longitude and latitude coordinates of the machine nest address to be screened. Represents the Gaussian kernel function with independent variable x, which is specifically expressed as , where is pi, and exp( ) represents the exponential function with the natural constant e as the base.

[0037] Similarly, the kernel density of the residential area corresponding to the machine nest address to be screened is normalized to obtain the second preferred address of the machine nest address to be screened, wherein the normalization method can be processed using the maximum and minimum value normalization method.

[0038] Step S203: determining a third preferred index for each to-be-screened machine nest address based on the POI kernel density of the area where each to-be-screened machine nest address is located.

[0039] Specifically, the POI (Point of Interest) kernel density refers to the concentration of various types of public infrastructure or stores. In this embodiment, the POI kernel density corresponding to the address of the machine nest to be screened is also calculated by the Gaussian kernel function. That is, each public infrastructure or store can be used as a POI location area, that is, an area of ​​interest. Furthermore, the POI kernel density corresponding to the address of the machine nest to be screened can be expressed by the formula:

[0040] in, Indicates the POI kernel density corresponding to the machine nest address to be screened, Indicates the total number of POI location areas, Represents the longitude and latitude coordinates of the particle in the i-th POI location area.

[0041] Similarly, the POI kernel density corresponding to the to-be-screened machine nest address is normalized to obtain the third preferred address of the to-be-screened machine nest address, wherein the normalization method may be processed using a maximum and minimum value normalization method.

[0042] Step S204: determining a fourth preferred index for each to-be-screened machine nest address based on the NIMBY index of the area where each to-be-screened machine nest address is located.

[0043] In this embodiment, the NIMBY index characterizes the distribution of residents who are disturbed by the machine nest in the vicinity of the machine nest address to be screened. The higher the NIMBY index, the fewer the nearby residents who are disturbed by the machine nest. Specifically, the calculation method of the NIMBY index corresponding to the machine nest address to be screened can be expressed as:

[0044] in, Indicates the NIMBY index corresponding to the address of the machine nest to be screened. Indicates the number of population statistics grids within the 200m buffer zone of the machine nest address to be screened. Indicates The population count of each demographic grid, Indicates The area where the demographic raster intersects with the 200-meter buffer of the current candidate address.

[0045] It should be noted that NIMBY (Not In My Backyard) refers to residents' concern that construction projects (such as garbage dumps, nuclear power plants, funeral homes and other NIMBY facilities) will have many negative impacts on their health, environmental quality and asset value, which breeds the mentality of "Don't build it in my backyard".

[0046] Similarly, the NIMBY index corresponding to the to-be-screened machine nest address is normalized to obtain the third preferred address of the to-be-screened machine nest address, wherein the normalization method may be a maximum-minimum normalization method.

[0047] Step S205: Based on a preset weight value, weighted sum is performed on the first preferred index, the second preferred index, the third preferred index and the fourth preferred index to determine a comprehensive evaluation index for each machine nest address to be screened.

[0048] In this embodiment, the preferred indicators obtained through feature analysis calculation in each data dimension respectively reflect the feature distribution of the machine nest address to be screened in the corresponding data dimension, and then by fusing the feature analysis results of all dimensions, the comprehensive evaluation results of the machine nest address to be screened can be more comprehensively reflected. The factors considered are more comprehensive, and the evaluation accuracy of the comprehensive evaluation results is higher.

[0049] Specifically, this embodiment fuses the feature analysis results under all dimensions by weighted summation. More specifically, in this embodiment, the weight value corresponding to the first preferred indicator is 0.4, the weight value corresponding to the second preferred indicator is 0.3, the weight value corresponding to the third preferred indicator is 0.2, and the weight value corresponding to the fourth preferred indicator is 0.1. It can be understood that the cumulative sum of the weight values ​​corresponding to all preferred indicators is 1, and the larger the weight, the greater the degree of attention to the feature analysis results of the dimension, and the implementer can set it according to the specific implementation scenario.

[0050] It should be noted that a comprehensive evaluation map can be constructed based on the comprehensive evaluation indicators corresponding to each nest address to be screened, and the missing values ​​can be set to 0. In the geographic information system (GIS), the "point to raster" tool can be used to convert all grid points into 100m×100m raster data, and the evaluation value of the point can be used as the attribute value of the raster. Select the appropriate color strip to render the raster data, select the appropriate base map, legend and scale for typesetting, and output the comprehensive evaluation map of the drone nest site selection according to the relevant map standards as a phased result for manual analysis and decision-making.

[0051] Step S300: Screening the machine nest addresses to be screened to obtain candidate addresses according to the comprehensive evaluation index and the distance distribution between the machine nest addresses to be screened.

[0052] The larger the value of the comprehensive evaluation index corresponding to the address of the machine nest to be screened, the more suitable the current address of the machine nest to be screened is as the site selection location of the drone nest from multiple perspectives. Based on this, the address of the machine nest to be screened can be filtered by the comprehensive evaluation index, and the part with a larger comprehensive evaluation result can be screened out as the alternative address.

[0053] Specifically, the addresses of the machine nests to be screened corresponding to the comprehensive evaluation index greater than the preset evaluation threshold are selected as candidate addresses, and the Euclidean distance between the particles in the area where each two candidate addresses are located is within the preset distance range. The evaluation threshold value range is (0,1), and the implementer can set it according to the specific implementation scenario.

[0054] It should be noted that when filtering the addresses of the drone nests to be screened, it is also necessary to constrain the one-way flight range of the drone to have at least one other drone nest and the minimum distance between nests. Specifically, in terms of the map range, it is required that there must be at least one other drone nest within the one-way flight range of the drone to enable the transfer of drones between nests, thereby building a network of drone nest collaboration. In order to make the random distribution initialization more dispersed and in line with common sense, there is a minimum limit distance between drone nests.

[0055] The one-way flight range of the drone refers to the circular area with the particle at the candidate address as the center and the maximum flight distance of the drone as the radius. The minimum limit distance between drone nests means that the distance between any two drone nests cannot be less than v, and v is less than 2r. v is initialized to 3000m, and this value can be appropriately modified according to the drone model and mission requirements.

[0056] Among them, it should be noted that the drone nest with a certain area is idealized as a single particle, which represents the geographical location of the nest on the map. That is, each alternative address can be represented by a corresponding particle, and each particle can also be represented by the corresponding longitude and latitude coordinates. The particles corresponding to all alternative addresses constitute the nest particle set.

[0057] Step S400, obtaining an objective function according to the coverage of residential communities within the drone service range corresponding to each candidate address, the comprehensive evaluation index and quantity distribution of the candidate addresses.

[0058] Specifically, according to the layout requirements, the reward and penalty mechanism are integrated to construct the objective function, that is, the objective function includes reward items and penalty items, such as Figure 3 As shown, the method for constructing the objective function is implemented by steps S401 to S403.

[0059] Step S401, obtaining a reward item according to the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold.

[0060] Specifically, the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold is calculated, and the cumulative sum of the differences corresponding to all candidate addresses is weighted to obtain the reward item, which can be expressed as:

[0061] in, Indicates the reward item. Describe the comprehensive evaluation index corresponding to the t-th candidate address, Indicates the total number of candidate addresses. Indicates the evaluation threshold. Indicates the reward weight for selecting a high-scoring area corresponding to an alternative address, which can be set according to the usage scenario, drone model, and mission nature.

[0062] Step S402, determine the first penalty item based on the total number of all candidate addresses by weighting; obtain the second penalty item based on the coverage of the residential area within the drone service range corresponding to each candidate address; and determine the comprehensive penalty item based on the cumulative sum of the first penalty item and the second penalty item.

[0063] In this embodiment, the penalty mechanism of the objective function mainly analyzes two aspects of characteristics. The first aspect is to consider the deployment cost of the drone nest, and the second aspect is to consider the coverage of the drone nest relative to the residential area.

[0064] First, in terms of the map range, the number of drone nests is minimized to reduce the overall deployment cost. That is, the more the total number of nests, the higher the penalty of the objective function. Specifically, the first penalty term is determined based on the total number of all candidate addresses, which can be expressed as ; Among them, is the first penalty term, Indicates the total number of candidate addresses. It is the penalty weight of each alternative address corresponding to the machine nest in the first aspect, which can be set according to the usage scenario, drone model and mission nature.

[0065] Secondly, in terms of map range, the more uncovered residential area particles there are, the higher the penalty of the objective function.

[0066] First, it is necessary to analyze the coverage of residential areas within the service range of the drone site selection. Specifically, the circular area with the particle point of each candidate address as the center and the length of the drone working radius as the radius is the drone service range of each candidate address.

[0067] In the map range, the drone service range corresponding to all drone nest locations should cover as many residential community particles as possible (to achieve the highest possible efficiency).

[0068] The Euclidean distance between the mass point in the area where each residential community is located and the mass point of the alternative address is calculated. When the value of the Euclidean distance is greater than the radius of the drone service range, the corresponding residential community is not within the drone service range of the alternative address; when the value of the Euclidean distance is less than or equal to the radius of the drone service range, the corresponding residential community is within the drone service range of the alternative address, which can be specifically expressed as:

[0069]

[0070] in, represents the Euclidean distance between the sth candidate address and the kth residential area, Indicates the longitude and latitude coordinates of the particle corresponding to the sth candidate address, Represents the longitude and latitude coordinates of the particle point corresponding to the kth residential area. Represents the radius of the circular area corresponding to the drone service range of the sth candidate address.

[0071] Indicates whether the kth residential area is within the drone service range of the sth candidate address. When , it means that the kth residential area is within the drone service range of the sth candidate address. When , it means that the kth residential area is not within the drone service range of the sth alternative address.

[0072] Furthermore, the total number of residential areas that are not within the service range of the drone at the candidate address is weighted to obtain the second penalty term, which can be expressed as:

[0073] in, represents the second penalty term, n represents the number of residential communities, min() represents the function for finding the minimum value, The penalty weight for each uncovered residential area particle can be set according to the usage scenario, drone model and mission nature.

[0074] Step S403: determining the objective function based on the difference between the reward item and the comprehensive penalty item.

[0075] In this embodiment, according to the layout requirements, the reward and penalty mechanism are integrated, and the objective function can be expressed as:

[0076] in, is the objective function, For reward items, is the first penalty term, As the second penalty term, when optimizing the objective function, it needs to be as large as possible to meet the requirements of the scenario.

[0077] Step S500, using the taboo search algorithm to optimize the objective function to obtain a nest site selection plan and the corresponding number of drone nests.

[0078] Tabu Search (TS) is a search method used to escape from local optimal solutions. The most important idea of ​​taboo search is to mark some objects corresponding to the local optimal solutions that have been searched, and try to avoid these objects in further iterative searches (rather than absolutely prohibiting loops), so as to ensure the exploration of different effective search paths. Taboo search involves neighborhoods, taboo tables, taboo lengths, candidate solutions, and contempt criteria. By optimizing the objective function through the taboo search algorithm, it is possible to avoid the objective function from falling into the local optimal solution, and to obtain the objective function with the best effect and a larger value. Finally, the optimal drone nest site selection plan, as well as the total number of drone nests in the corresponding plan and the maximized objective function value can be determined. The objective function value is used to measure the degree of optimization of the site selection plan. In addition, the coverage of the particles in each residential area is clarified, and the covered and uncovered community particles are identified.

[0079] As a specific example, this embodiment briefly introduces the process of optimizing the objective function using the taboo search algorithm. Figure 4 , which shows a flow chart of the steps for optimizing the objective function.

[0080] It should be noted that the terms in the objective function Z are not linearly related. This is a nonlinear, non-convex combinatorial optimization problem. Since Z involves multiple discrete decision variables, it means that multiple local optimal solutions may be encountered during the optimization process. According to the above characteristics of the objective function Z, traditional optimization methods are usually unable to effectively solve this problem.

[0081] The reason is that in non-convex optimization problems, traditional optimization methods can easily fall into local optimal solutions. Because these methods rely on gradients or other local information, they cannot effectively escape from local optimal points. In addition, the search space of combinatorial optimization problems is very large, especially when the number of potential machine nest locations and residential area particles is large, the number of possible solutions increases exponentially with the increase of locations and constraints.

[0082] In this embodiment, the taboo search algorithm is used to create a taboo table T to record and select the optimization process that has been performed. The algorithm marks the local optimal solution (taboo object) that has been solved in the past calculation process. In the subsequent iteration process, the local optimal solution (taboo object) stored in the taboo table is avoided, thereby obtaining more search areas and avoiding falling into the trap of the local optimal solution.

[0083] The taboo table T contains the following information: taboo objects, which represent the local optimal solutions that have been solved in the past calculation process. Taboo length TabuLength, which represents the number of taboo objects that can be stored in the taboo table, or the time that the taboo object remains in the taboo table.

[0084] The first step is to define the initial solution of the tabu search algorithm.

[0085] is the objective function Generate an initial solution , the local search will generate a neighborhood from this solution and perform subsequent adjacent searches. The objective function value of the initial solution is . Assume that the current solution , the optimal solution . Initial solution There are many ways to generate .

[0086] For example, the random generation method of the initial solution, that is, randomly generating a feasible solution according to the constraints in step S100 of the problem. Randomly select a number of grid positions that meet the conditions as the initial machine nest positions. The random generation method may result in low quality of the initial solution and require more iterations, but the requirements are relatively simple and the generalization is good.

[0087] The greedy algorithm generation method of the initial solution uses a greedy algorithm to first select several grid positions with the highest scores as the machine nest positions, and gradually construct an initial solution. Specifically, the grid position corresponding to the maximum value of the comprehensive evaluation index is used as the machine nest position, and an initial solution is gradually constructed. The flow chart is as follows: Figure 5 shown.

[0088] Specifically, initialize the initial solution to empty , indicating that the initial nest position set is an empty set. Initialize the remaining candidate position set to . Calculate all candidate locations The utility function The utility function is calculated based on the score of the grid point and its contribution to the mass point covering the residential area.

[0089]

[0090] Select the candidate location with the highest utility Add initial solution , . Update the initial solution, that is, the update process can be expressed as .

[0091] Update the candidate location set , delete all the positions related to the current selection For conflicting grid points (i.e. grid points whose distance is less than v), the update process can be expressed as:

[0092] in, Indicates the difference between the two, which reflects the candidate positions when the deletion conflict occurs. Check all remaining candidate positions Is the current solution satisfied? The above formula indicates that the candidate position set Update to the original candidate location set The difference between the set of locations that conflict with the set of locations that conflict.

[0093] If yes, continue to select the candidate position with the highest utility to add to the initial solution and update the candidate position set Otherwise, stop selecting. When it is empty or no new position can be selected without violating the constraints, the greedy selection process is stopped and the initial solution is obtained. The quality of the initial solution generated by the greedy algorithm is higher, which saves the number of iterations, but it is more likely to fall into the local optimal solution.

[0094] The second step is to initialize the taboo table. Initialize to an empty set: .

[0095] is the maximum capacity of the taboo table, that is, the maximum number of solutions allowed to be stored in the taboo table. The taboo length is a hyperparameter. If the length of the taboo table is too short, the effect of the taboo may be insufficient, and the algorithm may easily return to the solutions that have been visited; if it is too long, the flexibility of the algorithm may be limited, resulting in a decrease in search efficiency. This value is usually selected based on the scale of the community and the machine nest and the size of the urban space.

[0096] Let the maximum number of iterations be , which determines how many times the algorithm can perform neighborhood searches. More iterations usually mean a deeper search, which may make it easier to find the global optimal solution. But it also means that the algorithm requires more computing time. This value is usually selected based on the size of the community and machine nest and the size of the urban space. Initialize the number of iterations . Amnesty conditions This means that when the solution is in the taboo list and the objective function value of the solution is better than the current optimal objective function value, the solution can be considered as a taboo solution. The solution that meets the amnesty condition will be re-added to the taboo list and updated as the optimal solution.

[0097] The third step is iterative neighborhood generation and search.

[0098] Define the current solution Neighborhood set of , the neighborhood set is composed of the current solution The set of new solutions obtained by performing multiple random adjustments (movement, addition, deletion, and stationary) under constraints on the random number of nest positions in can be expressed as:

[0099] For each candidate solution , calculate the objective function value respectively . Select a candidate solution with the largest objective function value , but it is not necessarily the final solution.

[0100] Step 4: Check the taboo list and amnesty conditions.

[0101] Select the largest candidate solution from the neighborhood of the previous step When , check whether the solution has been stored in the taboo table T, expressed as:

[0102] when hour, , represents the maximum candidate solution If it is not in the taboo table, the maximum candidate solution Add to the taboo table. Further, when the maximum candidate solution When it is the optimal solution, update the optimal solution.

[0103] when hour, , represents the maximum candidate solution In the taboo table, the maximum candidate solution It is considered a taboo solution and cannot be accepted immediately. It needs to be further checked to see if it meets the amnesty conditions. If the amnesty conditions are met, the maximum candidate solution is re-added to the taboo table and updated as the optimal solution; if it does not meet the requirement, the maximum candidate solution is discarded. , select the suboptimal solution in the neighborhood and recheck it in the taboo table.

[0104] Step 4: Update the taboo list.

[0105] The new solution Add to taboo list If the length of the taboo table is Exceed , then delete the earliest solution in the taboo table.

[0106] Step 5: Update the optimal solution.

[0107] If the new solution The objective function value is better than the current optimal objective function value , replace the optimal solution with a new solution.

[0108] Step 6: Update the iteration status.

[0109] Update the current solution to the new solution , and increment the iteration count: . Indicates that the current solution Update to new solution , Indicates the number of iterations plus one.

[0110] When the maximum number of iterations is reached Or when the termination condition is met, the optimal solution and its corresponding maximum objective function value are output. The termination condition is adjusted according to the scenario, such as when the optimization objective function gain is less than the threshold, the loop phenomenon of the taboo table, and external termination, etc.

[0111] In the optimized drone nest site selection scheme, the output results determine the optimized nest site selection scheme and the total number of drone nests under the optimal site selection scheme. The maximized objective function value is output, and its value is used to measure the degree of optimization of the site selection scheme. In addition, the coverage of each residential community particle is clarified, and the covered and uncovered community particles are identified. Based on the above output results, the system can further adjust the optimization parameters and recalculate to provide a feasible solution to the drone nest site selection problem.

[0112] In summary, the site selection evaluation system of the present invention is based on multi-dimensional factors: A method for evaluating the site selection of drone nests with population coverage, community core density, POI core density, and NIMBY index as preferred indicators is proposed, and a systematic and standardized evaluation system is established. A method for optimizing the global layout of the entire region or city from a macro perspective is proposed, and the reasonable distribution and connectivity of drone nests in the region are ensured by quantifying the layout problem as an objective function. Site selection evaluation and alternative address charts are automatically generated through GIS technology and algorithms, which reduces manual intervention and improves the efficiency and accuracy of the site selection process. For the nonlinear and non-convex combinatorial optimization problem of drone nest site selection, a taboo search algorithm is introduced, and the taboo table management is used to avoid falling into the local optimal solution, thereby more effectively exploring a wider solution space and providing a more efficient solution. By adjusting the weights of the site selection evaluation indicators according to the use scenarios and purposes of the drones, the present invention provides a flexible site selection decision-making mechanism, which can adjust the site selection strategy according to actual needs and scenarios.

[0113] The embodiments described above are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, a person skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for selecting and arranging drone nests for urban management, characterized in that: The method comprises the following steps: Divide the target area to obtain the address of each machine nest to be screened, and the location of the machine nest to be screened meets the preset restriction conditions; According to the population coverage, residential community concentration, various infrastructure concentrations and the distribution of residents affected by the machine nest in the machine nest address to be screened, the machine nest address to be screened is evaluated to obtain a comprehensive evaluation index; According to the comprehensive evaluation index and the distance distribution between the addresses of the machine nests to be screened, the addresses of the machine nests to be screened are screened to obtain candidate addresses; The objective function is obtained based on the coverage of residential areas within the drone service range corresponding to each candidate address, the comprehensive evaluation index and quantity distribution of the candidate addresses; The taboo search algorithm is used to optimize the objective function to obtain the nest site selection plan and the corresponding number of UAV nests.

2. The method for selecting and arranging the site of drone nests for urban management according to claim 1 is characterized in that: The method for obtaining the address of the machine nest to be screened specifically includes: The target area is evenly divided into a dot grid with a preset length. A square area of ​​a preset size with each grid point on the dot grid as the center is the address of each drone nest. The drone nest address corresponding to the preset restriction condition is used as the nest address to be screened.

3. The method for selecting and arranging the site of drone nests for urban management according to claim 1 is characterized in that: The evaluation of the machine nest address to be screened is performed to obtain a comprehensive evaluation index, which specifically includes: Determine the first optimization index of each machine nest address to be screened based on the population coverage rate in the area where each machine nest address to be screened is located; Determine the second optimal index of each to-be-screened machine nest address based on the distribution density of residential communities in the area where each to-be-screened machine nest address is located; Determine the third optimization index of each to-be-screened machine nest address based on the POI kernel density of the area where each to-be-screened machine nest address is located; Determine the fourth optimization index of each machine nest address to be screened based on the NIMBY index of the area where each machine nest address to be screened is located; Based on the preset weight value, the first preferred index, the second preferred index, the third preferred index and the fourth preferred index are weightedly summed to determine the comprehensive evaluation index of each machine nest address to be screened, and the first preferred index, the second preferred index, the third preferred index and the fourth preferred index are all normalized values.

4. The method for selecting and arranging the site of drone nests for urban management according to claim 1 is characterized in that: The method of screening the machine nest addresses to be screened to obtain candidate addresses according to the comprehensive evaluation index and the distance distribution between the machine nest addresses to be screened specifically includes: The addresses of the machine nests to be screened corresponding to the comprehensive evaluation index being greater than the preset evaluation threshold are taken as candidate addresses, and the Euclidean distance between the particles in the areas where each two candidate addresses are located is within a preset distance range.

5. The method for selecting and arranging the site of drone nests for urban management according to claim 4 is characterized in that: The objective function is obtained according to the coverage of the residential area within the drone service range corresponding to each candidate address, the comprehensive evaluation index and quantity distribution of the candidate addresses, and specifically includes: Obtaining a reward item according to the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold; A first penalty item is determined based on the total number of all candidate addresses by weighting; a second penalty item is obtained based on the coverage of residential communities within the drone service range corresponding to each candidate address; and a comprehensive penalty item is determined based on the cumulative sum of the first penalty item and the second penalty item; The objective function is determined based on the difference between the reward term and the comprehensive penalty term.

6. The method for selecting and arranging the site of drone nests for urban management according to claim 5 is characterized in that: The reward item is obtained according to the difference between the comprehensive evaluation index of each candidate address and the evaluation threshold, specifically including: The difference between the comprehensive evaluation index of each candidate address and the evaluation threshold is calculated, and the cumulative sum of the differences corresponding to all candidate addresses is weighted to obtain a reward item.

7. The method for selecting and arranging the site of drone nests for urban management according to claim 5 is characterized in that: The second penalty item is obtained according to the coverage of the residential area within the service range of the drone corresponding to each candidate address, specifically including: A circular area with the mass point of each candidate address as the center and the length of the drone working radius as the radius is used as the drone service range of each candidate address; Calculate the Euclidean distance between the mass point in the area where each residential community is located and the mass point of the alternative address. When the value of the Euclidean distance is greater than the radius of the drone service range, the corresponding residential community is not within the drone service range of the alternative address; when the value of the Euclidean distance is less than or equal to the radius of the drone service range, the corresponding residential community is within the drone service range of the alternative address. The second penalty term is obtained by weighting the total number of residential communities that are not within the service range of the drone at the alternative address.

Images