Logistics unmanned aerial vehicle urban transportation network planning method based on transfer site selection

Abstract

The invention discloses a logistics unmanned aerial vehicle urban transportation network planning method based on transfer point site selection, and belongs to the field of unmanned aerial vehicle transportation network planning, and the method comprises the steps: carrying out the three-dimensional discretization modeling of a logistics unmanned aerial vehicle operation environment through employing a grid method based on ground space layout and urban airspace characteristics, and recognizing the potential obstacles of an airspace; designing a logistics unmanned aerial vehicle transportation network system architecture containing one network, two layers and three nodes; establishing a double-layer multi-target planning model including transfer point site selection and network construction by considering the number of transfer points, the operation cost and the balance of distribution network service pressure; and establishing a double-layer solving framework based on the self-organizing mapping neural network and a hybrid simulated annealing algorithm, and outputting a logistics unmanned aerial vehicle three-dimensional transportation network topological structure. According to the method, the logistics unmanned aerial vehicle three-dimensional transportation network is constructed for urban dense distribution requirements, the urban large-scale logistics unmanned aerial vehicles can be guided to fly orderly, and optimal configuration of low-altitude airspace resources is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of drone transportation network planning, and in particular to a logistics drone urban transportation network planning method based on transfer point site selection. Background Art

[0002] The problems of traffic congestion and inefficiency caused by the traditional delivery mode of logistics vehicles combined with couriers are becoming increasingly prominent, which not only restricts the timeliness of logistics delivery, but also increases the pressure of logistics delivery on the ground, bringing challenges to the sustainable development of cities. Drones, with their flexible and fast advantages, have brought innovative development ideas to the logistics and distribution industry. The continuous innovation of science and technology and the growing user demand indicate that drone logistics is expected to achieve more extensive commercial applications. However, large-scale logistics drone flights in cities in the future also pose challenges to urban air traffic management. The urban airspace is complex and the distribution points are dense. The contradiction between limited low-altitude resources and huge drone airspace demand is becoming increasingly prominent, which to a certain extent restricts the development of the drone logistics industry. Facing the multi-level urban spatial layout and high-density drone operation needs, it is particularly important to plan a suitable transportation network, which helps to guide drones to fly in an orderly and stable manner. Existing related studies are mainly based on urban spatial layout, proposing a macro-structure planning scheme for drone transportation networks, which have problems such as large network scale, dense route intersections, and uneven segment flow distribution. In addition, network topology operation safety and user demand distribution characteristics are rarely considered in transportation network design. Summary of the invention

[0003] The purpose of the present invention is to provide a logistics drone urban transportation network planning method based on transfer point site selection. Aiming at the dense urban distribution needs, taking into account the uneven distribution of distribution points and the complex low-altitude airspace environment in the actual application of drone logistics, the concept of transfer points is introduced into the transportation network based on the cargo transportation information of the starting and ending points, and the process of departure from the distribution point and reception at the demand point is separated to build a safe, efficient and standardized logistics drone transportation network, thereby accelerating the integration of drones into the urban airspace system.

[0004] To achieve the above object, the present invention provides a method for planning a logistics drone urban transportation network based on transfer point site selection, comprising the following steps:

[0005] Step 1: Based on the ground space layout and urban airspace characteristics, the grid method is used to perform three-dimensional discretization modeling of the logistics drone operation environment to identify potential airspace obstacles;

[0006] Step 2: Based on the delivery mode of logistics drones from the supply end to the user end, a logistics drone transportation network architecture with one network, two layers, and three nodes is designed;

[0007] Step 3: Considering the number of transfer points, operating costs, and the balance of distribution network service pressure, a two-layer multi-objective planning model including transfer point location selection and network construction is established;

[0008] Step 4: Build a two-layer solution framework based on the self-organizing map neural network and the hybrid simulated annealing algorithm to output the topological structure of the logistics drone three-dimensional transportation network.

[0009] Preferably, in step 1, building elevation data is obtained by remote sensing technology, and combined with information on no-fly zones, restricted zones, and dangerous zones, a number of cubic grids are used to fill the three-dimensional geographic space, and the airspace grids are divided into free grids and obstacle grids according to whether there are running obstacles in the grids. If there are no obstacles in the grids, it is a free grid, and if there are obstacles in the grids, it is an obstacle grid. The free grid is the flight area of ​​the logistics drone;

[0010] The center of each free grid is the potential location of the transfer point. Suppose the set of distribution points in the urban area is S, the set of demand points is R, and the set of potential locations of transfer points is T.

[0011] Preferably, in step 2, the logistics drone transportation network is designed with systematicity and hierarchy as the core, forming a structured logistics drone transportation system, by integrating the transfer layer network with the distribution layer network, realizing the connectivity between the distribution points, transfer points, and demand points, and forming an overall architecture of one network-two layers-three nodes;

[0012] The logistics drone takes off after loading the goods at the delivery end, flies according to the route planned in the logistics drone transportation network, lands and unloads the goods after arriving at the destination user end. According to the specific needs of the delivery task, after completing a delivery task, the logistics drone can choose to continue to fly to the next user end for delivery, or directly return to the starting delivery end for charging and maintenance.

[0013] Preferably, one network is the entire logistics drone transportation network, covering the entire process from the starting point to the end point;

[0014] The double-layer logistics drone transportation network is a layered structure, which includes a transfer layer network and a distribution layer network. The flight altitude layer where the transfer layer network is located is higher than the flight altitude layer where the distribution layer network is located. The transfer layer network is used to connect the transfer point and the distribution point. The route of the transfer layer network is called the transfer route. The distribution layer is used to connect the transfer point and the demand point. The route of the distribution layer is called the distribution route.

[0015] The three nodes are nodes in the logistics drone transportation network, including the distribution point, the transfer point and the demand point. The distribution point is the starting point of the goods, the transfer point is the hub in the transportation network, and the demand point is the end point of the logistics task.

[0016] Preferably, in step 3, based on the overall architecture of the logistics drone transportation network, a two-layer planning model of the logistics drone transportation network is established. The two-layer planning model of the logistics drone transportation network includes an upper model and a lower model. The upper model is a transfer point location model. The objective function of the transfer point location model includes the number of transfer points N t and delivery cost C;

[0017] For a potential location of a transfer point t(t∈T), define a binary variable x t Indicates whether it is enabled, expressed as:

[0018]

[0019] The number of transfer points in the transportation network N t It is expressed as:

[0020]

[0021] For an enabled transit point t, define a binary variable x tr Indicates whether it serves the demand point r (r∈R), expressed as:

[0022]

[0023] Considering the cargo volume as the user's demand weight, the distribution cost C of the distribution layer network is expressed as:

[0024]

[0025] In the formula, c tr is the amount of cargo delivered by drones from transfer point t to demand point r, d tr is the route distance between the transshipment point t and the demand point r.

[0026] Preferably, in the process of selecting the location of the transfer point, environmental restrictions and operational requirements are combined, with airspace restriction constraints, transfer point quantity constraints, distribution relationship constraints, service scope constraints, and transfer point pressure constraints as the constraints;

[0027] The airspace restriction constraint is expressed as:

[0028]

[0029] In the formula, g t is a binary variable, g t =1 means there is an obstacle at the potential location t of the transfer point, g t =0 means there is no obstacle at the location and the drone can pass;

[0030] The transfer point quantity constraint is expressed as:

[0031] N t≤N r ;

[0032] Number of transfer points N t Less than or equal to the number of demand points N r ;

[0033] The distribution relationship constraints are expressed as:

[0034]

[0035] Each demand point r is served by a transfer point;

[0036] The service range is used to define the demand points covered by the transfer point. The service range constraint is expressed as:

[0037]

[0038] In the formula, r u The service radius of the transfer point;

[0039] The pressure of the transfer point is defined as the total amount of goods that need to be delivered to the demand point at this point, which is less than the upper limit of the service pressure p max , the transfer point pressure constraint is expressed as:

[0040]

[0041] Preferably, the lower model is a network construction model, which incorporates route service pressure and length characteristics. The objective function of the network construction model includes the route service pressure standard deviation δ, characteristic path length L n , total length of transportation network D n Three functions;

[0042] Standard deviation of route service pressure δ:

[0043] According to the upper model, the service relationship between the transfer point and the demand point is obtained, and the selected transfer point set is defined as T u (T u ∈T), the transport layer network node set V T =T u ∪S, the transshipment route is l ij (i,j∈V T , i≠j), i and j represent the elements in the transport layer network node set, and the two nodes are not the same node. When the logistics drone transports goods, the shortest path is selected by default to perform the delivery task. The shortest path from the delivery point s (s∈S) to the demand point r (r∈R) is L sr , define the binary variable Indicates route l ij Is it on the shortest path L sr Inside, it is expressed as:

[0044]

[0045] The standard deviation of route service pressure is used to measure the difference in service pressure of transshipment routes. ij The service pressure is defined as the number of drone sorties passing through the route after completing all delivery tasks. ij , expressed as:

[0046]

[0047] In the formula, p sr The number of drone flights required to transport goods from distribution point s to demand point r, c sr is the amount of goods transported from the distribution point s to the demand point r, w u The maximum payload of the drone;

[0048] Average service pressure on transshipment routes It is expressed as:

[0049]

[0050] Where N l The number of transshipment routes;

[0051] At this time, the standard deviation δ of service pressure of all transshipment routes is expressed as:

[0052]

[0053] In the formula, k ij is a binary variable, k ij =1 means route l ij Enabled, k ij =0 means not enabled;

[0054] Characteristic path length L n :

[0055] In the logistics drone delivery network, the shortest path from the delivery point s to the demand point r is called the characteristic path. The characteristic path depends on the topological structure of the distribution network. The characteristic path length L of the distribution network is n It is defined as the average value of the characteristic path lengths from the distribution point to the demand point, expressed as:

[0056]

[0057] Where N s is the number of distribution points, d ij For transshipment route ij Length;

[0058] The total length of the transport network D n :

[0059] The network construction model is based on the total network length D n The minimum is the goal, expressed as:

[0060]

[0061] Preferably, in the process of constructing the logistics drone transportation network, the network topology and drone performance constraints are combined, with no isolated point constraints, transit number constraints, and drone range constraints as constraint conditions;

[0062] The logistics drone transportation network connects all the distribution points, transfer points and demand points to form a connected network. There are no isolated points in the network. The no isolated point constraint is expressed as:

[0063]

[0064] in, It means the continuous multiplication of the number of routes connected to each transport layer network node. It represents the continuous multiplication of the number of transit points served by each demand point;

[0065] The network is reasonably laid out by limiting the number of transfers. The transfer number constraint is expressed as:

[0066]

[0067] In the formula, Q max The maximum number of transfers allowed for logistics drones;

[0068] The UAV range constraint is expressed as:

[0069]

[0070] Where h is the height of the transport layer network, L u This is the maximum range of a logistics drone.

[0071] Preferably, in step four, in the two-layer planning model of the logistics drone transportation network, the upper layer is a two-target transfer point site selection model, and the lower layer is a three-target network construction model. The decision variables are all discrete variables, and there is a coupling relationship between the objective functions of the upper and lower models. The upper model selects a set of suitable transfer points in a limited spatial grid to match the service relationship with the demand points, which is regarded as a node clustering problem and is solved using a self-organizing mapping neural network. Through a competitive learning strategy between neurons, high-dimensional data is mapped to a low-dimensional topological space, and the demand point groups in the airspace are clustered to obtain a set of transfer points that meet the constraints and target conditions.

[0072] Preferably, the lower model plans appropriate routes between nodes to form a transportation network, incorporates the idea of ​​non-dominated sorting, and uses a hybrid simulated annealing algorithm to solve it. According to the cargo transportation information between the starting and ending points, a route library including transshipment routes and distribution routes is constructed, and the route characteristics are encoded to generate an initial population that meets the constraints. Through roulette wheel selection, crossover, and mutation, a child population is generated. After combining with the parent population, a non-dominated sorting strategy is used to generate a new generation of population. Through a double-layer iterative strategy, the upper-level transshipment point site selection and the lower-level network construction are continuously optimized until the entire system reaches a balanced state.

[0073] Therefore, the present invention adopts the above-mentioned logistics drone urban transportation network planning method based on transfer point site selection, which has the following beneficial effects:

[0074] The present invention comprehensively considers the urban spatial layout, network topology and UAV flight performance, adopts the grid method to discretize the UAV operation space modeling, designs a multi-level logistics UAV three-dimensional transportation network system architecture, combines the number of transfer points and distribution cost factors, takes into account the balance of transportation network service pressure, constructs a logistics UAV transportation network double-layer multi-objective planning model, and designs an algorithm framework based on self-organizing map neural network and hybrid genetic algorithm to solve the UAV transportation network structure, promotes the structured and refined development of the logistics UAV transportation network, and provides an efficient and intelligent solution for future urban logistics distribution.

[0075] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] Figure 1 A schematic diagram of a method flow chart of an embodiment of the present invention;

[0077] Figure 2 This is a schematic diagram of the urban transportation network architecture of a logistics drone according to an embodiment of the present invention;

[0078] Figure 3 This is a diagram of the operation scenario of the logistics drone urban transportation network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0079] In order to make the purpose, technical scheme and advantages disclosed in the embodiments of the present invention clearer, the embodiments of the present invention are further described in detail in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the embodiments of the present invention and are not intended to limit the embodiments of the present invention. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application. Examples of the embodiments are shown in the accompanying drawings, where the same or similar reference numerals throughout represent the same or similar elements or elements with the same or similar functions.

[0080] It should be noted that the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product or server that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units that are not explicitly listed or inherent to these processes, methods, products or devices.

[0081] Like reference numerals and letters denote similar items in the following drawings, and thus, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0082] In the description of the present invention, it should be noted that the terms "upper", "lower", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, or are directions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.

[0083] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "setting", "installation" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0084] Example

[0085] like Figure 1-3 As shown, the method for planning a logistics drone urban transportation network based on transfer point location selection described in the present invention includes the following steps:

[0086] Step 1: Based on the ground space layout and urban airspace characteristics, the grid method is used to conduct three-dimensional discrete modeling of the logistics drone operating environment to identify potential obstacles in the airspace.

[0087] The building elevation data is obtained through remote sensing technology, and combined with the information of no-fly zones, restricted zones, and dangerous zones, a number of cubic grids are used to fill the three-dimensional geographic space. Depending on whether there are obstacles to operation in the grid, the airspace grid is divided into free grids and obstacle grids. Logistics drones can only fly in free grids.

[0088] The center of each free grid is the potential location of the transfer point. Suppose the set of distribution points in the urban area is S, the set of demand points is R, and the set of potential locations of transfer points is T.

[0089] Step 2: Based on the delivery mode of logistics drones from the supply end to the user end, design a logistics drone transportation network architecture with one network, two layers, and three nodes.

[0090] After loading the goods at the delivery end, the logistics drone takes off and flies along the route planned in the transportation network. After reaching the destination user end, it lands and unloads the goods. According to the specific needs of the delivery task, after completing a delivery task, the logistics drone can choose to continue to fly to the next user end for delivery, or directly return to the starting delivery end for charging and maintenance.

[0091] The logistics drone transportation network takes systematicity and hierarchy as the core of design, forming a structured logistics drone transportation system. By integrating the transfer layer network with the distribution layer network, the connection between distribution points, transfer points and demand points is realized, forming an overall architecture of one network, two layers and three nodes.

[0092] One network: refers to the entire logistics drone transportation network, covering the entire process from the starting point to the end point, ensuring the seamless connection of the entire logistics drone delivery process;

[0093] Double-layer: The logistics drone transportation network is a layered structure, including two layers: the transfer layer network and the distribution layer network. The transfer layer network is located at a higher flight altitude layer, connecting the transfer point and the distribution point. The route of this layer is called the transfer route. The distribution layer network is located at a lower flight altitude layer, connecting the transfer point and the demand point. The route of this layer is called the distribution route.

[0094] Three nodes: The nodes in the transportation network are divided into distribution points, transshipment points and demand points. The distribution point is the starting point of the goods. The transshipment point is the hub in the transportation network and is the bridge between the distribution point and the demand point. The demand point is the end point of the logistics task and marks the final delivery of the goods.

[0095] Step 3: Considering the number of transfer points, operating costs, and the balance of distribution network service pressure, a two-layer multi-objective planning model including transfer point site selection and network construction is established.

[0096] Based on the overall architecture of the logistics drone transportation network, a two-layer planning model for the logistics drone transportation network is established. The upper model is the transfer point location model, and the objective function includes two aspects:

[0097] (1) Number of transfer points N t

[0098] When building a logistics drone transportation network, the number of transfer points is directly related to the operating cost and air traffic management efficiency. For the potential location of the transfer point t (t∈T), define a binary variable x t Indicates whether it is enabled, expressed as:

[0099]

[0100] At this time, the number of transfer points in the transportation network is N t It can be expressed as:

[0101]

[0102] (2) Delivery cost C

[0103] The distance between the transfer point and the demand point is a key factor affecting the distribution cost. Usually, there is a positive relationship between the two. For the enabled transfer point t, define a binary variable x tr Indicates whether it serves the demand point r (r∈R), expressed as:

[0104]

[0105] Considering that the quantity of goods is also an important factor in the distribution cost, it can be regarded as the user's demand weight, which in turn affects the distribution strategy and cost calculation. Therefore, the distribution cost C of the distribution layer network can be expressed as:

[0106]

[0107] In the formula, c tr is the amount of cargo delivered by drones from transfer point t to demand point r, d tr is the route distance between the transshipment point t and the demand point r;

[0108] In the process of selecting the site for the transfer point, the following constraints need to be considered in combination with environmental restrictions and operational requirements:

[0109] (1) Airspace restrictions: There are many obstacles in the low-altitude airspace of cities, such as buildings and signal blind spots. Logistics drones are strictly prohibited from passing through, and transfer points are also restricted from being located in these obstacle areas, which can be expressed as:

[0110]

[0111] In the formula, g t is a binary variable, g t =1 means that there is an obstacle at the potential location t of the transfer point, and this location cannot be selected. t =0 means there are no obstacles at the location and the drone can pass;

[0112] (2) Constraints on the number of transfer points. The setting of transfer points should ensure the rationality of the distribution service coverage, while also preventing the dispersion of resources due to their excessive number. The number of transfer points should not exceed the number of demand points N. r , expressed as:

[0113] N t ≤N r ;

[0114] (3) Distribution relationship constraints: In order to ensure the specificity of the service, each demand point is served by only one transfer point, which can be expressed as:

[0115]

[0116] (4) Service scope constraint: To ensure the accessibility of the distribution service, the service scope is used to define the demand points that the transfer point can cover, which can be expressed as:

[0117]

[0118] In the formula, r u The service radius of the transfer point;

[0119] (5) Transshipment point pressure constraint: A transshipment point can serve more than one demand point. If it undertakes too many delivery tasks, it will lead to overload operation. The transshipment point pressure is defined as the total amount of goods that need to be delivered to the demand points at this point, which needs to be less than the service pressure upper limit p. max , expressed as:

[0120]

[0121] According to the results of the transshipment point location selection and layout, the lower model is a network construction model, which incorporates the route service pressure and length characteristics in order to plan a convenient transportation network topology. The objective function includes three aspects:

[0122] (1) Standard deviation of route service pressure δ

[0123] According to the upper model, the service relationship between the transfer point and the demand point can be obtained. The selected transfer point set is defined as T u (T u∈T), at this time, the transport layer network node set V T =T u ∪S, the transshipment route is l ij (i,j∈V T , i≠j), i and j represent the elements in the transport layer network node set, and the two nodes are not the same node. When the logistics drone transports goods, the shortest path is selected by default to perform the delivery task. The shortest path from the delivery point s (s∈S) to the demand point r (r∈R) is L sr , define the binary variable Indicates route l ij Is it on the shortest path L sr Inside, it is expressed as:

[0124]

[0125] A transshipment point can be connected to multiple transshipment routes. If a transshipment route bears too much distribution demand, the service pressure of the transshipment route will be unbalanced, which will affect the stability of the entire network. Therefore, the standard deviation of route service pressure is used to measure the difference in transshipment route service pressure. ij The service pressure is defined as the number of drone sorties passing through the route after completing all delivery tasks. ij , expressed as:

[0126]

[0127] In the formula, p sr The number of drone flights required to transport goods from distribution point s to demand point r, c sr is the amount of goods transported from the distribution point s to the demand point r, w u The maximum payload of the drone;

[0128] Average service pressure on transshipment routes It can be expressed as:

[0129]

[0130] Where N l The number of transshipment routes;

[0131] At this time, the standard deviation of service pressure of all transshipment routes δ can be expressed as:

[0132]

[0133] In the formula, k ij is a binary variable, k ij =1 means route l ij Enabled, k ij =0 means not enabled;

[0134] (2) Characteristic path length L n

[0135] In the logistics drone delivery network, there may be more than one path from the delivery point s to the demand point r. The shortest path is also called the characteristic path. The characteristic path depends on the topological structure of the distribution network. Its length is directly related to the flight efficiency and energy consumption of the drone. The characteristic path length L of the distribution network is n It is defined as the average value of the characteristic path lengths from the distribution point to the demand point, expressed as:

[0136]

[0137] Where N s is the number of distribution points, d ij For transshipment route ij Length;

[0138] (3) Total length of the transportation network D n

[0139] The total length of the transport network directly affects the cost and pressure of network management. If the total length of the network is too large, it may lead to longer transport time and lower service efficiency. At the same time, the complexity of the network structure will also increase, requiring more resources and more sophisticated scheduling to ensure stable operation of the network. Therefore, the network construction model is based on the total length of the network D. n The minimum is the goal, expressed as:

[0140]

[0141] During network construction, the following constraints need to be considered in combination with the network topology and drone performance limitations:

[0142] (1) No isolated point constraints. The transportation network needs to connect all distribution points, transfer points and demand points to form a connected network. There must be no isolated points in the network, which can be expressed as:

[0143]

[0144] in, It means the continuous multiplication of the number of routes connected to each transport layer network node. It represents the continuous multiplication of the number of transit points served by each demand point;

[0145] (2) Transit times constraint: The characteristic path from the distribution point s to the demand point r may pass through multiple nodes for transit. In order to avoid the increase of logistics costs due to frequent transit, the network is reasonably laid out by limiting the transit times, which can be expressed as:

[0146]

[0147] In the formula, Q max The maximum number of transfers allowed for logistics drones;

[0148] (3) UAV range constraint: The flight path of the logistics UAV from the delivery point s to the demand point r needs to strictly comply with its maximum range limit to ensure that the UAV can complete the delivery task safely and efficiently within its endurance range, which can be expressed as:

[0149]

[0150] Where h is the height of the transport layer network, L u This is the maximum range of a logistics drone.

[0151] Step 4: Build a two-layer solution framework based on the self-organizing map neural network and the hybrid simulated annealing algorithm to output the topological structure of the logistics drone three-dimensional transportation network.

[0152] In the two-layer planning model of the logistics drone transportation network, the upper layer is a two-objective transfer point location model, and the lower layer is a three-objective network construction model. The decision variables are all discrete variables, and the objective functions of the upper and lower models are coupled, which is an NP-hard problem. The essence of the upper model is to select a set of suitable transfer points in a limited airspace grid to match the service relationship with the demand points. It can be regarded as a node clustering problem and is solved by a self-organizing map neural network. The algorithm maps high-dimensional data to a low-dimensional topological space through a competitive learning strategy between neurons, and can cluster the demand point groups in the airspace to obtain a set of transfer points that meet the constraints and target conditions.

[0153] In order to meet the objective function of the transfer point location model, the distribution cost between the neuron α and the input vector X is calculated, and the neuron with the lowest distribution cost is defined as the activated neuron. The activated neuron is taken as the center, and the weights of the neurons in the activated neuron and its neighborhood are updated until the predetermined number of training times is reached. Since the model constrains the number of transfer points not to exceed the number of demand points, in order to find a transfer point set that meets the dual objectives of minimizing the number of transfer points and the distribution cost, the algorithm cluster number is set to the range of [1, N r ], for the location selection results of the transfer points under each cluster number, if the model constraints are met, the solution is retained. According to the dominant relationship between feasible solutions, a set of Pareto optimal solutions can be obtained. There is no distinction between good and bad solutions in the solution set. Based on the objective function value, the scoring method is used to select the solution with the highest score as the final solution of the upper model;

[0154] The lower model aims to plan appropriate routes between nodes to form a transportation network, incorporates the idea of ​​non-dominated sorting, and uses a hybrid simulated annealing algorithm to solve it. According to the cargo transportation information between the starting and ending points, a route library containing transshipment routes and distribution routes is constructed. The route characteristics are encoded to generate an initial population that meets the constraints. Roulette wheel selection, crossover, and mutation are used to generate child populations. After combining with the parent population, a non-dominated sorting strategy is used to generate a new generation of populations. Through a double-layer iterative strategy, the upper-level transshipment point location and the lower-level network construction are continuously optimized until the entire system reaches a state of equilibrium.

[0155] Therefore, the logistics drone urban transportation network planning method based on transfer point selection described in the present invention constructs a three-dimensional transportation network for logistics drones in response to the city's intensive distribution needs, which helps to guide the orderly flight of large-scale logistics drones in the city and achieve the optimal allocation of low-altitude airspace resources.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A logistics drone urban transportation network planning method based on transfer point location selection, characterized by: The following steps are involved: Step 1: Based on the ground space layout and urban airspace characteristics, the grid method is used to perform three-dimensional discretization modeling of the logistics drone operation environment to identify potential airspace obstacles; Step 2: Based on the delivery mode of logistics drones from the supply end to the user end, a logistics drone transportation network architecture with one network, two layers, and three nodes is designed; Step 3: Considering the number of transfer points, operating costs, and the balance of distribution network service pressure, a two-layer multi-objective planning model including transfer point location selection and network construction is established; Step 4: Build a two-layer solution framework based on the self-organizing map neural network and the hybrid simulated annealing algorithm to output the topological structure of the logistics drone three-dimensional transportation network.

2. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 1 is characterized by: In step 1, the building elevation data is obtained through remote sensing technology, and combined with the information of no-fly zones, restricted areas, and dangerous areas in the air, a number of cubic grids are used to fill the three-dimensional geographic space. According to whether there are running obstacles in the grid, the airspace grid is divided into free grids and obstacle grids. If there are no obstacles in the grid, it is a free grid, and if there are obstacles in the grid, it is an obstacle grid. The free grid is the flight area of ​​the logistics drone; The center of each free grid is the potential location of the transfer point. Suppose the set of distribution points in the urban area is S, the set of demand points is R, and the set of potential locations of transfer points is T.

3. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 2 is characterized in that: In step 2, the logistics drone transportation network takes systematicity and hierarchy as the core of design, forming a structured logistics drone transportation system. By integrating the transfer layer network with the distribution layer network, the connectivity between distribution points, transfer points and demand points is achieved, forming an overall architecture of one network, two layers and three nodes; The logistics drone takes off after loading the goods at the delivery end, flies according to the route planned in the logistics drone transportation network, lands and unloads the goods after arriving at the destination user end. According to the specific needs of the delivery task, after completing a delivery task, the logistics drone can choose to continue to fly to the next user end for delivery, or directly return to the starting delivery end for charging and maintenance.

4. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 3 is characterized by: One network is the entire logistics drone transportation network, covering the entire process from starting point to destination; The double-layer logistics drone transportation network is a layered structure, which includes a transfer layer network and a distribution layer network. The flight altitude layer where the transfer layer network is located is higher than the flight altitude layer where the distribution layer network is located. The transfer layer network is used to connect the transfer point and the distribution point. The route of the transfer layer network is called the transfer route. The distribution layer is used to connect the transfer point and the demand point. The route of the distribution layer is called the distribution route. The three nodes are nodes in the logistics drone transportation network, including the distribution point, the transfer point and the demand point. The distribution point is the starting point of the goods, the transfer point is the hub in the transportation network, and the demand point is the end point of the logistics task.

5. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 4 is characterized in that: In step 3, based on the overall architecture of the logistics drone transportation network, a two-level planning model of the logistics drone transportation network is established. The two-level planning model of the logistics drone transportation network includes an upper model and a lower model. The upper model is a transfer point location model. The objective function of the transfer point location model includes the number of transfer points N t and delivery cost C; For a potential location of a transfer point t(t∈T), define a binary variable x t Indicates whether it is enabled, expressed as: The number of transfer points in the transportation network N t It is expressed as: For an enabled transit point t, define a binary variable x tr Indicates whether it serves the demand point r (r∈R), expressed as: Considering the cargo volume as the user's demand weight, the distribution cost C of the distribution layer network is expressed as: In the formula, c tr is the amount of cargo delivered by drones from transfer point t to demand point r, d tr is the route distance between the transshipment point t and the demand point r.

6. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 5 is characterized by: In the process of selecting the location of the transfer point, environmental restrictions and operational requirements are combined, with airspace restriction constraints, transfer point quantity constraints, distribution relationship constraints, service scope constraints, and transfer point pressure constraints as the constraints; The airspace restriction constraint is expressed as: In the formula, g t is a binary variable, g t =1 means there is an obstacle at the potential location t of the transfer point, g t =0 means there is no obstacle at the location and the drone can pass; The transfer point quantity constraint is expressed as: N t ≤N r ; Number of transfer points N t Less than or equal to the number of demand points N r ; The distribution relationship constraints are expressed as: Each demand point r is served by a transfer point; The service range is used to define the demand points covered by the transfer point. The service range constraint is expressed as: In the formula, r u The service radius of the transfer point; The pressure of the transfer point is defined as the total amount of goods that need to be delivered to the demand point at this point, which is less than the upper limit of the service pressure p max , the transfer point pressure constraint is expressed as:

7. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 6 is characterized by: The lower model is a network construction model, which incorporates route service pressure and length characteristics. The objective function of the network construction model includes the route service pressure standard deviation δ, characteristic path length L n , total length of transportation network D n Three functions; Standard deviation of route service pressure δ: According to the upper model, the service relationship between the transfer point and the demand point is obtained, and the selected transfer point set is defined as T u (T u ∈T), the transport layer network node set V T =T u ∪S, the transshipment route is l ij (i,j∈V T , i≠j), i and j represent the elements in the transport layer network node set, and the two nodes are not the same node. When the logistics drone transports goods, the shortest path is selected by default to perform the delivery task. The shortest path from the delivery point s (s∈S) to the demand point r (r∈R) is L sr , define the binary variable Indicates route l ij Is it on the shortest path L sr Inside, it is expressed as: The standard deviation of route service pressure is used to measure the difference in service pressure of transshipment routes. ij The service pressure is defined as the number of drone sorties passing through the route after completing all delivery tasks. ij , expressed as: In the formula, p sr The number of drone flights required to transport goods from distribution point s to demand point r, c sr is the amount of goods transported from the distribution point s to the demand point r, w u The maximum payload of the drone; Average service pressure on transshipment routes It is expressed as: Where N l The number of transshipment routes; At this time, the standard deviation δ of service pressure of all transshipment routes is expressed as: In the formula, k ij is a binary variable, k ij =1 means route l ij Enabled, k ij =0 means not enabled; Characteristic path length L n : In the logistics drone delivery network, the shortest path from the delivery point s to the demand point r is called the characteristic path. The characteristic path depends on the topological structure of the distribution network. The characteristic path length L of the distribution network is n It is defined as the average of all characteristic path lengths from the distribution point to the demand point, expressed as: Where N s is the number of distribution points, d ij For transshipment route ij Length; The total length of the transport network D n : The network construction model is based on the total network length D n The minimum is the goal, expressed as:

8. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 7 is characterized by: In the process of constructing the logistics drone transportation network, the network topology structure and drone performance constraints are combined, with no isolated point constraints, transfer number constraints, and drone range constraints as constraints; The logistics drone transportation network connects all the distribution points, transfer points and demand points to form a connected network. There are no isolated points in the network. The no isolated point constraint is expressed as: in, It means the continuous multiplication of the number of routes connected to each transport layer network node. It represents the continuous multiplication of the number of transit points served by each demand point; The network is reasonably laid out by limiting the number of transfers. The transfer number constraint is expressed as: In the formula, Q max The maximum number of transfers allowed for logistics drones; The UAV range constraint is expressed as: Where h is the height of the transport layer network, L u This is the maximum range of a logistics drone.

9. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 8 is characterized by: In step four, in the two-layer planning model of the logistics drone transportation network, the upper layer is a two-target transfer point site selection model, and the lower layer is a three-target network construction model. The decision variables are all discrete variables, and there is a coupling relationship between the objective functions of the upper and lower models. The upper model selects a set of suitable transfer points in a limited spatial grid to match the service relationship with the demand points. It is regarded as a node clustering problem and is solved using a self-organizing mapping neural network. Through the competitive learning strategy between neurons, the high-dimensional data is mapped to a low-dimensional topological space, and the demand point groups in the airspace are clustered to obtain a set of transfer points that meet the constraints and target conditions.

10. The method for planning a logistics drone urban transportation network based on transfer point location selection according to claim 9 is characterized in that: The lower model plans appropriate routes between nodes to form a transportation network, incorporates the idea of ​​non-dominated sorting, and uses a hybrid simulated annealing algorithm to solve it. According to the cargo transportation information between the starting and ending points, a route library containing transshipment routes and distribution routes is constructed. The route characteristics are encoded to generate an initial population that meets the constraints. Roulette wheel selection, crossover, and mutation are used to generate child populations. After combining with the parent population, a non-dominated sorting strategy is used to generate a new generation of populations. Through a double-layer iterative strategy, the upper-level transshipment point location and the lower-level network construction are continuously optimized until the entire system reaches a balanced state.

Images