Unmanned aerial vehicle logistics distribution route determination method and device

Abstract

The invention discloses an unmanned aerial vehicle logistics distribution route determination method and device, the unmanned aerial vehicle logistics distribution route determination method is executed in a computing device, and the method comprises the steps: obtaining a distribution task list, and recording the distribution information of at least one distribution task in the distribution task list, the distribution information comprises a distribution address, a distribution weight and the latest distribution time; arranging the delivery addresses in sequence to generate a plurality of initial delivery routes; taking minimization of total energy consumption, maximization of task completion rate and satisfaction of the latest delivery time window as optimization objectives, and constructing a route optimization objective function; based on a route optimization objective function, performing optimization screening on the initial delivery route to obtain an optimized candidate route; and performing obstacle avoidance and smoothing processing on each candidate route in a three-dimensional space by using an A star algorithm, thereby outputting a final delivery route.

Description

Methods and devices for determining drone logistics delivery routes Technical Field

[0001] This invention relates to the field of drone logistics delivery, and specifically to a method and apparatus for determining drone logistics delivery routes. Background Technology

[0002] In recent years, with the deep integration of drone technology and the logistics industry, drone delivery has become an important part of the smart logistics system and is regarded as an effective means to solve the "last mile" delivery problem. However, existing drone route planning technology still faces severe challenges in actual large-scale and commercial logistics applications, and its core planning algorithm is difficult to meet the multi-dimensional and dynamic constraints required in complex real-world scenarios.

[0003] Currently, most mainstream UAV path planning technologies originate from the field of robotics. Their research focuses on planning a collision-free, shortest geometric path from start to finish for a single UAV to perform a single task in a known or partially known 3D environment. Traditional algorithms, such as Dijkstra's algorithm, A* (A* algorithm) and its variants, can effectively find paths in static raster maps or point clouds, but their optimization objectives are usually only path length or flight time, lacking in-depth modeling of logistics and delivery business logic. While sampling methods based on Rapid Exploratory Random Trees (RRT) have probabilistic completeness in high-dimensional spaces, the generated paths are often highly random and lack smoothness, and similarly do not consider upper-level business requirements.

[0004] The limitations of these traditional methods are amplified dramatically in real-world logistics scenarios, specifically: First, they suffer from a single planning dimension. Existing methods are generally designed for short-range, point-to-point flights and static path generation, failing to handle multi-point delivery tasks involving multiple customers. In other words, they cannot intelligently plan the globally optimal access sequence and path for multiple dispersed customer points within a single flight. Second, they lack business constraints. The planning process does not incorporate key logistics operational elements, such as delivery time requirements for different packages, differences in customer service priorities, and the dynamic impact of package weight and volume on drone endurance and maneuverability. Algorithms focus only on spatial obstacle avoidance, ignoring these time and business constraints that determine delivery efficiency and service quality. Third, they lack dynamic adaptability. Real-world airspace environments are constantly changing, including temporary no-fly zones, sudden weather conditions, and real-time air traffic control instructions. Traditional static path planning methods cannot achieve online replanning to ensure flight safety and timeliness.

[0005] Therefore, there is an urgent need to build a new route planning methodology that can simultaneously couple spatial geographic information, aircraft physical characteristics, package attributes, business rules and dynamic airspace information. This system needs to have the ability to perform global optimization and dynamic decision-making under multiple objectives and constraints, so as to truly support the efficient, reliable and intelligent operation of the UAV logistics network. Summary of the Invention

[0006] In view of the above problems, the present invention is proposed to provide a method and apparatus for determining unmanned aerial vehicle (UAV) logistics delivery routes to overcome or at least partially solve the above problems.

[0007] According to one aspect of the present invention, a method for determining unmanned aerial vehicle (UAV) logistics delivery routes is provided, executed in a computing device. The method includes: acquiring a delivery task list, wherein the delivery task list records delivery information for at least one delivery task, the delivery information including delivery address, delivery weight, and latest delivery time; sequentially arranging the delivery addresses to generate multiple initial delivery routes; constructing a route optimization objective function with the optimization objectives of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window, wherein minimizing total energy consumption is calculated based on the distance between delivery addresses and the delivery weight, and the time window constraint is achieved by determining whether the expected time of arrival at the target delivery address is earlier than its latest delivery time; optimizing and filtering the initial delivery routes based on the route optimization objective function to obtain optimized candidate routes; and using the A* algorithm to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route.

[0008] Optionally, the method for determining unmanned aerial vehicle (UAV) logistics delivery routes according to the present invention further includes: constructing a set of nodes based on each delivery address, and constructing an edge set between the elements in the node set, wherein the node set and the edge set together define a directed graph structure; determining the distance and compliance coefficient between delivery addresses corresponding to each edge, wherein the compliance coefficient indicates whether the path between delivery addresses crosses a no-fly zone; determining the energy consumption and the flight time required by the UAV based on the distance, delivery weight, wind speed parameters, and required climb altitude; and associating the distance, compliance coefficient, energy consumption, and flight time as edge attributes of the corresponding edges into the directed graph structure.

[0009] Optionally, in the method for determining unmanned aerial vehicle (UAV) logistics delivery routes according to the present invention, the delivery addresses are arranged sequentially to generate multiple initial delivery routes, including: performing a full permutation of all node numbers in the directed graph structure to generate multiple number sequences; inserting the starting point number at the beginning of each number sequence and the return point number at the end, thereby obtaining each initial delivery route.

[0010] Optionally, in the method for determining unmanned aerial vehicle (UAV) logistics delivery routes according to the present invention, energy consumption is expressed by the following formula: ;in, For delivery address Energy consumption in the room For delivery address The distance between, For the delivery address The delivery weight at the time of takeoff. For delivery address The headwind speed of the drone For delivery address The drone climbed to a higher altitude. This represents the weight value.

[0011] Optionally, in the UAV logistics delivery route determination method according to the present invention, the route optimization objective function is expressed by the following formula: Among them, among them, The value represents the fitness of route optimization. For delivery address Energy consumption in the room To reach the delivery address Time, For delivery address The latest delivery time, This is a penalty for late completion; R represents the task completion rate: R = number of completed tasks / total number of tasks. This represents the weight value.

[0012] Optionally, in the method for determining unmanned aerial vehicle (UAV) logistics delivery routes according to the present invention, the initial delivery routes are optimized and screened based on the route optimization objective function to obtain optimized candidate routes, including: for each initial delivery route, using the route optimization objective function, calculating the fitness among the delivery addresses it contains, and determining the initial delivery route as a feasible delivery route when the fitness meets the threshold requirement; using a genetic algorithm to encode each feasible delivery route into chromosomes, and performing population evolution through selection, crossover, and mutation operators to obtain each candidate route.

[0013] Optionally, in the UAV logistics delivery route determination method according to the present invention, the A* algorithm is used to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route. This includes: constructing a spatial three-dimensional mesh model covering each delivery address, wherein the three-dimensional mesh model contains obstacle areas; determining the cost function of the A* algorithm based on the straight-line distance between delivery addresses, the cost per unit of wind resistance, the height difference between the UAV and the building roof during flight, and the penalty factor for approaching no-fly zones, to obtain an improved A* algorithm; using the improved A* algorithm, combined with candidate routes, performing route search in the spatial three-dimensional mesh model to generate a three-dimensional delivery route; and using Bezier or B-spline curves to smooth the three-dimensional delivery route to obtain the final delivery route.

[0014] Optionally, in the UAV logistics delivery route determination method according to the present invention, the cost function is expressed by the following formula: ;in, Let cost function be The straight-line distance between delivery addresses. Cost per unit of wind resistance The height difference between the drone and the building roof during flight. A penalty factor for drones approaching no-fly zones during flight. , , , All are weight values.

[0015] Optionally, the method for determining unmanned aerial vehicle (UAV) logistics delivery routes according to the present invention further includes the steps of: determining whether the delivery point corresponding to each delivery address meets the delivery conditions; replacing the delivery address that does not meet the delivery conditions with the nearest delivery address, and updating the directed graph structure.

[0016] Optionally, in the UAV logistics delivery route determination method according to the present invention, the delivery conditions include at least one of the following conditions: whether the delivery point exists within a 2*2m radius. 2 The above-mentioned unobstructed horizontal platform; whether the tilt angle of the landing area of ​​the delivery point is less than 10°; whether the number of people, heat sources and electromagnetic interference within the landing radius of the delivery point exceeds the standard.

[0017] According to another aspect of the present invention, a drone logistics delivery route determination device is provided, residing in a computing device. The device includes: an acquisition module, adapted to acquire a delivery task list, wherein the delivery task list records delivery information of at least one delivery task, the delivery information including delivery address, delivery weight, and latest delivery time; a sorting module, adapted to sequentially arrange the delivery addresses to generate multiple initial delivery routes; a construction module, adapted to construct a route optimization objective function with the optimization objectives of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window, wherein minimizing total energy consumption is calculated based on the distance between delivery addresses and the delivery weight, and the time window constraint is achieved by determining whether the expected time of arrival at the target delivery address is earlier than its latest delivery time; a filtering module, adapted to optimize and filter the initial delivery routes based on the route optimization objective function to obtain optimized candidate routes; and a processing module, adapted to use the A* algorithm to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route.

[0018] According to another aspect of the present invention, a computing device is provided, comprising: at least one processor; and a memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, the program instructions including instructions for performing the methods described above.

[0019] According to another aspect of the present invention, a readable storage medium storing program instructions is provided, which, when read and executed by a computing device, causes the computing device to perform the method described above.

[0020] According to the present invention, a single drone can complete multiple delivery points in a single takeoff and landing, intelligently generating globally optimal 3D flight routes. This elevates traditional "point-to-point" planning to highly efficient "multi-point roving" planning, significantly improving single-drone delivery efficiency and operational economy. Furthermore, this solution deeply integrates logistics business rules and flight physics constraints, simultaneously considering multiple factors such as delivery time windows, the dynamic impact of package weight on battery power, and the drone's endurance limits during planning. Therefore, it generates not just a simple "shortest path," but an "optimal task path" under multiple constraints including time, battery power, and business rules, ensuring the timeliness and feasibility of delivery tasks. In addition, by integrating real-time updated 3D geographic information and airspace regulations, this solution ensures that the generated flight routes always meet obstacle avoidance and airspace compliance requirements and can respond to dynamic changes (such as temporary no-fly zones) with online replanning. This fundamentally improves the safety and reliability of drone logistics delivery, laying a solid technical foundation for large-scale commercial applications.

[0021] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0022] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: FIG1 shows a block diagram of the physical components (i.e., hardware) of a computing device 100; FIG2 shows a flowchart of a drone logistics delivery route determination method 200 according to an embodiment of the present invention; FIG3 shows a structural schematic diagram of a drone logistics delivery route determination device 300 according to an embodiment of the present invention. Detailed Implementation

[0023] Exemplary embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure may be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the disclosure to those skilled in the art.

[0024] Existing drone route planning technologies mostly focus on single tasks, short-range flights, and static path generation, which cannot meet the complex constraints of real-world logistics scenarios such as "multi-point delivery, time-limited delivery, dynamic airspace changes, and limited power".

[0025] Traditional path planning methods, such as those based on Dijkstra, A*, or RRT, can generate obstacle avoidance paths in three-dimensional space, but they lack the ability to model logistics operations, such as the impact of package weight on range.

[0026] To address the problems existing in the prior art, this invention proposes a method for determining unmanned aerial vehicle (UAV) logistics delivery routes. This method enables a single UAV to complete multiple delivery points in a single takeoff and landing, intelligently generating globally optimal three-dimensional routes. It elevates traditional "point-to-point" planning to efficient "multi-point roving" planning, significantly improving single-unit delivery efficiency and operational economy. Furthermore, this solution deeply integrates logistics business rules and flight physics constraints, simultaneously considering multiple factors such as delivery time windows, the dynamic impact of package weight on battery power, and the UAV's endurance limits during planning. Therefore, it generates not just a simple "shortest path," but an "optimal task path" under multiple constraints including time, battery power, and business rules, ensuring the timeliness and feasibility of delivery tasks. In addition, by integrating real-time updated three-dimensional geographic information and airspace regulations data, this solution ensures that the generated routes always meet obstacle avoidance and airspace compliance requirements and can respond to dynamic changes (such as temporary no-fly zones) with online replanning. This fundamentally improves the safety and reliability of UAV logistics delivery, laying a solid technical foundation for large-scale commercial applications.

[0027] The method for determining drone logistics delivery routes can be executed in computing device 100. Computing device 100 can be implemented as a server, such as an application server, web server, etc.; it can also be implemented as a desktop computer, laptop computer, processor chip, tablet computer, etc., but is not limited thereto. Figure 1 shows a block diagram of the physical components (i.e., hardware) of computing device 100. In a basic configuration, computing device 100 includes at least one processing unit 102 and system memory 104. According to one aspect, depending on the configuration and type of computing device, system memory 104 includes, but is not limited to, volatile memory (e.g., random access memory), non-volatile memory (e.g., read-only memory), flash memory, or any combination of such memory.

[0028] According to one aspect, system memory 104 includes operating system 105. System memory 104 also includes program module 150. According to one aspect, operating system 105, for example, is adapted to control the operation of computing device 100. Furthermore, examples are practiced in conjunction with graphics libraries, other operating systems, or any other applications, and are not limited to any particular application or system. This basic configuration is shown in Figure 1 by the components within dashed lines 108. According to one aspect, computing device 100 has additional features or functions. For example, according to one aspect, computing device 100 includes additional data storage devices (removable and / or non-removable), such as disks, optical discs, or magnetic tapes. Such additional storage is shown in Figure 1 by removable storage device 109 and non-removable storage device 110.

[0029] As stated above, according to one aspect, a program module 150 is stored in system memory 104. According to one aspect, the program module 150 can be implemented as one or more computer program products. This application does not limit the type of computer program product, and may include, for example, email, word processing applications, spreadsheet applications, database applications, slideshow applications, drawing or computer-aided applications, web browsers, etc. In some embodiments according to this application, computer programs / instructions related to the UAV logistics delivery route determination method 200 are encapsulated as a computer program product. When these computer programs / instructions are executed by a processor (i.e., processing unit 102), the UAV logistics delivery route determination method 200 according to this application is implemented.

[0030] According to one aspect, examples can be practiced in circuits comprising discrete electronic components, in packages or integrated electronic chips containing logic gates, in circuits utilizing microprocessors, or on a single chip containing electronic components or a microprocessor. For example, examples can be practiced via a system-on-a-chip (SOC) in which each or more components shown in FIG. 1 can be integrated on a single integrated circuit. According to one aspect, such an SOC device may include one or more processing units, graphics units, communication units, system virtualization units, and various application functions, all integrated (or “burned in”) as a single integrated circuit onto a chip substrate. When operating via the SOC, the functions described herein can be operated via dedicated logic integrated on a single integrated circuit (chip) with other components of the computing device 100. Embodiments of the invention can also be practiced using other techniques capable of performing logical operations (e.g., AND, OR, and NOT), including but not limited to mechanical, optical, fluid, and quantum technologies. Additionally, embodiments of the invention can be practiced within a general-purpose computer or in any other circuit or system.

[0031] According to one aspect, the computing device 100 may also have one or more input devices 112, such as a keyboard, mouse, pen, voice input device, touch input device, etc. It may also include output devices 114, such as a monitor, speaker, printer, etc. The foregoing devices are examples and other devices may also be used. The computing device 100 may include one or more communication connections 116 that allow communication with other computing devices 118, such as printing devices, like printers. Examples of suitable communication connections 116 include, but are not limited to: RF transmitters, receivers, and / or transceiver circuitry; Universal Serial Bus (USB); parallel and / or serial ports.

[0032] As used herein, the term computer-readable medium includes computer storage medium. Computer storage medium can include volatile and non-volatile, removable and non-removable media implemented using any method or technology for storing information (e.g., computer-readable instructions, data structures, or program modules). System memory 104, removable storage device 109, and non-removable storage device 110 are examples of computer storage media (i.e., memory storage). Computer storage media can include random access memory (RAM), read-only memory (ROM), electrically erasable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital universal disc (DVD) or other optical storage, magnetic tape, magnetic tape, disk storage or other magnetic storage devices, or any other article of manufacture that can be used to store information and is accessible by computer device 100. According to one aspect, any such computer storage medium can be part of computing device 100. Computer storage media does not include carrier waves or other transmitted data signals.

[0033] According to one aspect, a communication medium is implemented by computer-readable instructions, data structures, program modules, or other data in a modulated data signal (e.g., a carrier wave or other transmission mechanism), and includes any information transmission medium. According to one aspect, the term "modulated data signal" describes a signal having one or more sets of characteristics or altered in a manner that encodes information in the signal. By way of example and not limitation, a communication medium includes wired media such as wired networks or direct wired connections, and wireless media such as acoustic, radio frequency (RF), infrared, and other wireless media.

[0034] In an embodiment of the invention, a computing device 100 is configured to execute a drone logistics delivery route determination method 200 according to the invention. The computing device 100 includes one or more processors and one or more readable storage media storing program instructions that, when configured to be executed by the one or more processors, cause the computing device to execute the drone logistics delivery route determination method 200 of the present invention.

[0035] Figure 2 shows a flowchart of a drone logistics delivery route determination method 200 according to an embodiment of the present invention. Method 200 is adapted to be executed in a computing device (e.g., the aforementioned computing device 100).

[0036] As shown in Figure 2, the purpose of Method 200 is to realize a method for determining UAV logistics delivery routes that can meet the composite constraints of "multi-point delivery, time-limited delivery, dynamic airspace changes, and power constraints".

[0037] Method 200 begins with step 202, in which a delivery task list is obtained. The delivery task list records delivery information t_i for at least one delivery task. The delivery information includes delivery address A_i, delivery weight w_i, and latest delivery time L_i.

[0038] Specifically, the computing device can obtain a list of delivery tasks from the upstream order management platform through the system interface. This list can be a structured data set that fully describes all delivery tasks that need to be completed within the current planning period.

[0039] The core of the delivery task list is the detailed information of at least one delivery task it contains. Each task is abstracted as a data object, and its key attributes (i.e., delivery information) include at least: Delivery address: This is the spatial target of the task, usually provided in the form of high-precision geographic coordinates (such as latitude and longitude), and is the basis for achieving accurate route planning and last-mile delivery. This address directly corresponds to a node in the subsequent directed graph structure of the path.

[0040] Delivery weight: This attribute represents the physical mass of the package. It is not isolated business data, but is directly input into the drone's high-precision energy consumption model, becoming a key variable in calculating flight drag and motor power consumption. For example, a heavier package will significantly increase the drone's energy consumption per unit distance, thus directly affecting its maximum range and route selection.

[0041] Latest delivery time: This attribute imbues the task with a sense of urgency, constituting a "time window" constraint in route planning. The computing device must ensure that the drone arrives at the address no later than this limit when planning the route; otherwise, a late delivery penalty will be incurred, directly impacting service quality and optimization goals.

[0042] In addition, in practical applications, delivery information may also include other attributes that affect route planning, such as task priority (used to determine the execution order in the event of resource conflicts) and special handling requirements for packages (such as temperature-controlled transportation). This embodiment does not limit these attributes.

[0043] In a specific example, suppose a drone delivery center receives a delivery task list from a smart city dispatch platform, containing three urgent tasks: Task T001: Deliver an emergency medical test kit to the rooftop helipad of the emergency center at the Municipal People's Hospital. The specific delivery information is as follows: the delivery address is located at longitude 118.12345°, latitude 39.56789°; the total weight of the test kit and its insulated box is 2.1 kg; given the urgency of the medical supplies, the latest delivery time is set to 10:30 AM today.

[0044] Task T002: Deliver a smartphone ordered from an e-commerce platform to a smart parcel locker in a residential community. Delivery information: Delivery address is 118.13570° longitude, 39.57890° latitude; package weight is light, only 0.4 kg; it is a standard delivery service, with the latest delivery time being 2 PM today.

[0045] Task T003: A box of urgently needed auto parts needs to be delivered to an auto repair shop on the outskirts of the city. The delivery information is as follows: the delivery address is at longitude 118.20012°, latitude 39.62234°; since the parts are metal, they weigh 4.8 kg; the factory requires them to arrive before noon for repairs, so the latest delivery time is 12:00 noon today.

[0046] Meanwhile, to achieve safe, compliant, and efficient route planning, the computing device, after acquiring the business-oriented delivery task list, will simultaneously access and integrate various key physical and regulatory environment data. These data collectively constitute a multi-dimensional decision space for route planning, ensuring that the generated routes are not only theoretically optimized but also physically feasible and safe to fly.

[0047] Physical environment data includes 3D geographic information data, building models, and no-fly zone boundaries.

[0048] Three-dimensional geographic information data is primarily acquired through digital surface models (DSMs). Unlike digital elevation models (DEMs), which only reflect ground elevation, DSMs include the height information of all objects on the ground surface (such as buildings and trees), thus realistically reflecting the three-dimensional terrain below the flight airspace. They are typically derived from satellite remote sensing, aerial photogrammetry, or LiDAR scans and are updated regularly. DSMs are the foundation for constructing a three-dimensional flyable space. Computing equipment uses DSMs to accurately identify obstacles above the ground (such as hills and mountains) and calculate the altitude differences between ascent and descent, which is crucial for estimating energy consumption and ensuring safe airspace clearance with the ground.

[0049] Building models are detailed 3D models of buildings in an urban environment, typically including the building's geometry, outline, height, and even facade structure (such as windows and balconies). They can be obtained through methods such as City Information Modeling (CIM), Building Information Modeling (BIM), or high-precision oblique photogrammetry 3D reality models. Building models are the core basis for providing beyond-line-of-sight obstacle avoidance. When drones fly in urban canyons, they must maintain a safe distance from facilities such as communication towers on the sides and roofs of buildings. Detailed building models enable algorithms to accurately plan detour paths in 3D space and can also be used to assess the feasibility of potential take-off and landing points on or around building rooftops.

[0050] A no-fly zone is an airspace area where drones are absolutely prohibited from entering, defined by a polygon or polyhedron. This typically includes airspace above sensitive areas such as airport clear zones, military control zones, government buildings, and nuclear power plants. It serves as a rigid safety constraint for flight path planning. When constructing flight paths, any boundary crossing the no-fly zone will incur extremely high penalties or be completely excluded, fundamentally eliminating the possibility of illegal flights.

[0051] The controlled environment data includes current wind speed fields and dynamic airspace restrictions.

[0052] The current wind field refers to real-time wind field data covering the planned area, containing both direction and magnitude vectors. This includes wind conditions at different altitudes, as wind speed and direction can differ significantly between the ground and the air. It can be obtained through weather forecasting service APIs, weather radar data, or a network of miniature weather sensors deployed throughout the city. It's easy to understand that flying against the wind significantly increases the drone's power consumption, while flying with the wind may reduce energy consumption. The wind field data is directly input into the power consumption model to more accurately predict the energy requirements for each flight segment.

[0053] Dynamic airspace restrictions refer to temporary and changeable airspace usage regulations. Examples include temporary no-fly zones established for large-scale events (such as sporting events or concerts), emergency airspace designated for public emergencies (such as fires or rescue operations), or real-time air traffic control instructions based on flow management. These can be obtained through the dynamic notification system of airspace management departments and real-time information released by unmanned aerial vehicle (UTM) traffic management platforms. Dynamic airspace restrictions enable the system to respond to changes. Unlike static no-fly zones, dynamic restrictions require that the planning algorithm is not a one-off event but rather capable of receiving real-time data streams and quickly triggering flight path replanning when new airspace restrictions are detected, ensuring continuous compliance.

[0054] In a specific example, suppose the computing device receives a delivery task that needs to traverse the city center. After obtaining the task list, the system immediately retrieves: DSM data for the area, which reveals a hilly region along the task path; a building model accurately shows the dense high-rise buildings on both sides of the flight path, forming an "urban canyon"; the no-fly zone boundary indicates a large airport near the flight path, whose airspace protection zone must be bypassed; the real-time wind field shows northerly winds that day, with a strong westerly jet stream at a height of 100 meters; and dynamic airspace restrictions indicate a temporary no-fly zone has been established due to a celebration event scheduled for the city center square during the planned delivery period.

[0055] By fusing and processing the aforementioned multi-source data, the system-generated flight routes can: automatically bypass airport airspace and temporary celebration areas; maintain legally required safe distances from building sidewalls when traversing urban canyons; calculate reasonable gradients to save energy when flying over hills; and ultimately plan an energy-efficient path that utilizes tailwinds as much as possible and avoids strong crosswinds. This fully demonstrates the powerful and robust planning capabilities of this invention when facing complex real-world constraints.

[0056] In some embodiments, after the computing device obtains the delivery task list, step 202 further includes the following sub-steps: First, construct a node set based on each delivery address, and construct an edge set between the elements in the node set, and the node set and the edge set together define a directed graph structure.

[0057] Specifically, the construction of the node set: Each delivery address (including the warehouse departure point, the delivery point for each package, and the return point after the mission) is abstracted as a node (Vertex) in the graph. The collection of all these nodes constitutes all the location points in the graph model. Nodes represent the spatial locations that the drone must visit or may pass through during the mission.

[0058] Construction of the edge set: An edge is created between any two distinct nodes in the node set. For example, the edge A->B from node A (warehouse) to node B (customer 1) is a separate edge from the edge B->A from B, because factors such as headwinds and hill climbs can lead to different round-trip flight costs. An edge represents a possible direct flight path between nodes. Initially, this is a fully connected or partially connected graph, representing all theoretically direct flight routes.

[0059] Then, determine the distance and compliance coefficient between the delivery addresses corresponding to each edge. The compliance coefficient indicates whether the path between delivery addresses crosses a no-fly zone.

[0060] Then, the energy consumption and the required flight time of the drone are determined based on the distance, delivery weight, wind speed parameters, and required altitude.

[0061] Finally, distance, compliance coefficient, energy consumption, and flight time are used as edge attributes of the corresponding edges and associated with the directed graph structure.

[0062] For example, the delivery task can be abstracted as a directed graph structure G = (V, E): V is the set of delivery points, including the starting point (delivery station) v_0, delivery points v_1~v_n, and return point v_r. E is the set of edges, which includes multiple edges e_{i,j}.

[0063] Each edge e_{i,j} has an attached attribute: distance. Energy consumption Time required Compliance coefficient (If the path crosses a no-fly zone, set it to ∞ or a high penalty value).

[0064] In some embodiments, energy consumption is expressed by the following formula: ;in, For delivery address Energy consumption in the room For delivery address The distance between, For the delivery address The delivery weight at the time of takeoff. For delivery address The headwind speed of the drone For delivery address The drone climbed to a higher altitude. This represents the weight value.

[0065] In some embodiments, method 200 further includes the following steps: determining whether the delivery point corresponding to each delivery address meets the delivery conditions, replacing the delivery address that does not meet the delivery conditions with the nearest delivery address, and updating the directed graph structure.

[0066] Delivery conditions include at least one of the following: Does the delivery point have a 2*2m area? 2 The above-mentioned unobstructed horizontal platform and the tilt angle of the landing area of ​​the delivery point are less than 10°. The number of people, heat sources and electromagnetic interference within the landing radius of the delivery point are not exceeded.

[0067] Subsequently, in step 204, the delivery addresses are arranged sequentially to generate multiple initial delivery routes.

[0068] Specifically, first, all node numbers in the directed graph structure are permuted to generate multiple number sequences. Then, the starting point number is inserted at the beginning of each number sequence, and the return point number is inserted at the end, thus obtaining each initial delivery route.

[0069] For example, node numbering: First, assign a unique numerical number to all delivery points (i.e., task delivery addresses) in the directed graph structure that need to be arranged in order. For example, if there are 5 delivery tasks, they would be numbered 1, 2, 3, 4, 5.

[0070] Note: The departure point (warehouse / distribution station) and return point are usually not involved in numbering and arrangement at this stage because their locations are fixed start and end points for each mission.

[0071] Generate all permutations: Perform all possible permutations of the above delivery point numbers. A permutation is the generation of all possible, non-repeating sequences. For n delivery points, there are n! (n factorial) different permutations. For example, for 3 delivery points (numbered 1, 2, 3), the permutations can generate the following 6 different sequences: [1, 2, 3], [1, 3, 2], [2, 1, 3], [2, 3, 1], [3, 1, 2], [3, 2, 1]. Each permutation represents a unique assumption about the task execution order. By generating a large number of permutations, sufficient diversity of starting points for subsequent optimization processes is ensured, which is an important prerequisite for evolutionary algorithms to find high-quality global solutions.

[0072] Subsequently, a complete flight route is constructed, combining fixed origin and destination points with variable mission sequences to form a structurally complete delivery route.

[0073] Next, insert the starting point and return point. For each pure delivery point sequence generated in the previous step, insert the starting point number (e.g., number 0) at the beginning and the return point number (e.g., number r) at the end.

[0074] Taking the sequence [2, 3, 1] as an example, after insertion, it becomes the complete flight path: [0, 2, 3, 1, r].

[0075] This ensures that each considered path is a complete flight mission with a beginning and an end. It starts from the warehouse (0), visits the various customer points in a specified order (e.g., 2→3→1), and finally returns to the base (r).

[0076] Finally, the mapping from sequence to route: each numbered sequence like [0, 2, 3, 1, r] is defined as an initial delivery route. In the computational model, this sequence can be directly mapped back to the previously constructed weighted directed graph structure. Two adjacent numbers in the sequence correspond to a directed edge in the graph. For example, the flight path corresponding to the sequence [0, 2, 3, 1, r] is: edge (0->2) → edge (2->3) → edge (3->1) → edge (1->r). By querying the edge attributes (distance, energy consumption, time, etc.) associated with these edges in the graph, the total cost of completing this full route can be quickly calculated, thereby evaluating its merits.

[0077] The core advantage of this "permutation-insertion" method lies in its systematicity and unbiasedness. It does not rely on any prior experience, but rather provides a robust and comprehensive initial solution set for the optimization algorithm by exhaustively enumerating all possible combinations of task sequences. These initial routes constitute the first generation population in the genetic algorithm. Based on this, the algorithm will iteratively develop an optimal delivery solution in terms of energy consumption, time, and compliance by simulating natural evolutionary mechanisms such as "selection, crossover, and mutation."

[0078] After generating the initial delivery route, in step 206, a route optimization objective function is constructed with the optimization objectives of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window. The minimization of total energy consumption is calculated based on the distance between delivery addresses and the delivery weight, and the time window constraint is achieved by determining whether the expected time to arrive at the target delivery address is earlier than its latest delivery time.

[0079] In some embodiments, the route optimization objective function is expressed by the following formula: ;in, The value represents the fitness of route optimization. For delivery address Energy consumption in the room To reach the delivery address Time, For delivery address The latest delivery time, This is a penalty for late submissions. R represents the task completion rate, where R = number of completed tasks / total number of tasks. This represents the weight value.

[0080] Then, in step 208, the initial delivery routes are optimized and screened based on the route optimization objective function to obtain optimized candidate routes.

[0081] Specifically, firstly, for each initial delivery route, the fitness among the delivery addresses it contains is calculated using the route optimization objective function, and when the fitness meets the threshold requirement, the initial delivery route is determined to be a feasible delivery route.

[0082] Then, each of the feasible delivery routes is encoded into chromosomes using a genetic algorithm, and population evolution is performed through selection, crossover, and mutation operators to obtain each candidate route.

[0083] In a specific example, a genetic algorithm is used to optimize the drone logistics delivery route for multiple objectives. The core process is as follows: 1. Treat the full permutation sequence of each initial delivery route as a chromosome. Example: Chromosome [0,2, 3, 1, r] represents the delivery order as 0→2→3→1→r.

[0084] 2. Initial population generation: Randomly generate an initial population of 50-200 individuals, ensuring that each individual is a valid permutation, representing different delivery order schemes.

[0085] 3. Fitness assessment: Solve each permutation sequence using the above-mentioned route optimization objective function and calculate: F = α × total flight distance + β × time window violation penalty + γ × energy consumption over-limit penalty. The fitness value is defined as: 1 / (1 + F). Ensure that fitness is positively correlated with solution quality. Use this function to comprehensively evaluate the flight efficiency, timeliness, and energy consumption feasibility of each solution, and simultaneously verify the degree of compliance of the solution with endurance, power limits, and no-fly zone constraints.

[0086] 4. Selection operation: Based on fitness ratio, a roulette wheel selection method is used to select high-quality individuals and retain the best individuals in the current generation to directly enter the next generation, ensuring the convergence of the algorithm.

[0087] 5. Crossover operation: Apply the sequential crossover (OX) or partial mapping crossover (PMX) operator to generate new individuals by exchanging segments of the parent chromosome while maintaining the relative order of genes. Example: Crossover between the parent [1,2,3,4,5] and [3,5,4,1,2] may produce [1,5,4,2,3].

[0088] 6. Mutation operations: Implement exchange mutation or inversion mutation strategies to randomly adjust the gene sequence of individuals, introduce appropriate random perturbations, maintain population diversity, and prevent premature convergence.

[0089] 7. Feasibility repair: Detect and repair infeasible solutions that violate key constraints (such as power consumption and time windows), and ensure that all individuals meet the requirements of the actual problem through local adjustment or rearrangement mechanisms.

[0090] 8. Iterative evolution: Repeatedly execute the selection-crossover-mutation-repair evolutionary cycle, retaining the historical best solution in each generation to continuously optimize the population quality. The algorithm terminates after reaching the preset maximum number of iterations. 9. Result output: Output one or more near-optimal delivery sequence schemes, providing an efficient and feasible route planning solution for drone logistics delivery.

[0091] This method effectively solves the multi-constraint path optimization problem in drone logistics delivery by simulating the natural evolution process, and significantly improves planning efficiency while ensuring the quality of the solution.

[0092] Then, in step 210, the A* algorithm is used to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route.

[0093] Specifically, firstly, a spatial three-dimensional mesh model covering each delivery address is constructed, and the three-dimensional mesh model includes obstacle areas.

[0094] This step involves digitizing and discretizing the real flight environment to provide a computable spatial basis for path search.

[0095] Model Construction: Based on the acquired Digital Surface Model (DSM) and building models, the system constructs a three-dimensional mesh model (Voxel Grid) covering all delivery addresses and surrounding airspace. This model divides the continuous physical space into countless uniform small cubic units (voxels).

[0096] Obstacle Identification: In the model, each cube cell is labeled as either "free space" or "obstacle region." Obstacle regions primarily include solid parts of buildings, known permanent obstacles (such as high-voltage power line towers), and other spaces that the drone absolutely cannot traverse. This 3D mesh model serves as the search map for the A* algorithm, transforming the complex geometric obstacle avoidance problem into a search problem of finding connected paths within a discrete mesh.

[0097] Then, based on the straight-line distance between delivery addresses, the cost per unit of wind resistance, the height difference between the drone and the building roof during flight, and the penalty factor for approaching no-fly zones, the cost function of the A* algorithm is determined, resulting in the improved A* algorithm.

[0098] Traditional A* algorithms typically only consider distance as a cost. This invention introduces multiple cost factors to make key improvements, making it more suitable for the actual flight needs of logistics drones.

[0099] The new cost function is defined as: .in: -Basic distance cost: The straight-line distance from the current node to the next node. This is the basis for path shortestification. - Wind resistance affects cost: Based on real-time wind speed field data. The algorithm calculates the angle between the path direction and the wind direction, assigning a higher cost when flying against the wind, encouraging the algorithm to prioritize paths with lower wind resistance to achieve energy savings. - Flight altitude cost: This cost factor encourages drones to fly at safe and compliant altitudes. It penalizes ultra-low-altitude flights (high risk) and unnecessary ultra-high-altitude flights (high energy consumption) in building clusters. There is usually an "economic cruising altitude range" within which the algorithm tends to generate paths. - Airspace risk cost: This factor creates a "risk buffer zone" for no-fly zone boundaries and dynamically restricted areas. Even if the path does not directly cross the no-fly zone, it will be penalized if it gets too close to its boundary. This provides a safety redundancy for flight and helps to cope with GPS positioning errors. , , , All are weight values.

[0100] Subsequently, using the improved A-Star algorithm and candidate routes, a route search was performed in a three-dimensional spatial grid model to generate a three-dimensional delivery route.

[0101] Using the improved A* algorithm described above, a search is performed for each pair of consecutive delivery points in the candidate routes within the constructed 3D mesh model.

[0102] Search process: Starting from the initial point, the algorithm continuously explores the surrounding grid and always selects the node with the smallest total cost estimate (current cost + estimated cost to the target) for expansion until the target point is found. Since the cost function includes obstacle, wind resistance, altitude, and risk information, the path automatically generated by the algorithm is naturally a collision-free, energy-efficient, safe, and compliant 3D path.

[0103] The output of this step is a three-dimensional delivery route connecting two delivery points, consisting of a series of grid center points. This route is safe to fly, but may consist of sharp zigzag lines.

[0104] Finally, the 3D delivery route is smoothed using Bezier or B-spline curves to obtain the final delivery route.

[0105] It is worth noting that the polyline path generated directly by the A* algorithm has two problems: it does not conform to the flight mechanics characteristics of drones, the drone cannot change direction instantaneously, it may lead to a passenger-like flight experience, and it is not conducive to package safety and the lifespan of the aircraft.

[0106] Therefore, this embodiment uses B-spline curves or Bézier curves to fit and smooth the original polyline path. B-spline curves are preferred due to their local controllability and convexity preservation; adjusting a control point in the path will not affect the entire curve, facilitating local optimization. The final delivery route obtained after smoothing is a continuous, smooth, and continuously curvatured three-dimensional spatial curve. This route not only safely avoids obstacles but also conforms to the dynamic constraints of the UAV. After commands are issued, it can be stably and accurately tracked by the flight control system, ensuring the smoothness and reliability of the entire delivery process.

[0107] Through this series of steps, the present invention refines a macroscopic mission instruction into a high-quality three-dimensional flight trajectory that can be directly executed by the UAV, fully demonstrating the solution's excellent precision planning and safety assurance capabilities in complex urban environments.

[0108] The method provided by this invention enables a single drone to complete multiple delivery points in a single takeoff and landing. It intelligently generates globally optimal 3D flight paths, transforming traditional "point-to-point" planning into highly efficient "multi-point roving" planning, significantly improving single-drone delivery efficiency and operational economy. Furthermore, this solution deeply integrates logistics business rules and flight physics constraints, simultaneously considering multiple factors such as delivery time windows, the dynamic impact of package weight on battery power, and the drone's endurance limits during planning. This results not just a simple "shortest path," but an "optimal task path" under multiple constraints including time, battery power, and business rules, ensuring the timeliness and feasibility of delivery tasks. In addition, by integrating real-time updated 3D geographic information and airspace regulations, this solution ensures that the generated flight paths always meet obstacle avoidance and airspace compliance requirements and can respond to dynamic changes (such as temporary no-fly zones) with online replanning. This fundamentally improves the safety and reliability of drone logistics delivery, laying a solid technical foundation for large-scale commercial applications.

[0109] Figure 3 shows a schematic diagram of a drone logistics delivery route determination device 300 according to an embodiment of the present invention. The device 300 resides in a computing device. The device 300 includes an acquisition module 302, a sorting module 304, a construction module 306, a filtering module 308, and a processing module 310 that are coupled to each other.

[0110] The acquisition module 302 is adapted to acquire a delivery task list, which records delivery information for at least one delivery task, including delivery address, delivery weight, and latest delivery time. The sorting module 304 is adapted to sequentially arrange the delivery addresses to generate multiple initial delivery routes. The construction module 306 is adapted to construct a route optimization objective function with the optimization goals of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window. Minimizing total energy consumption is calculated based on the distance between delivery addresses and the delivery weight, and the time window constraint is achieved by determining whether the expected arrival time at the target delivery address is earlier than its latest delivery time. The filtering module 308 is adapted to optimize and filter the initial delivery routes based on the route optimization objective function to obtain optimized candidate routes. The processing module 310 is adapted to use the A* algorithm to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route.

[0111] It should be noted that the working principle and process of the device 300 provided in this embodiment are similar to those of the method 200 described above. For relevant details, please refer to the description of the method 200 described above, which will not be repeated here.

[0112] The various techniques described herein can be implemented in combination with hardware or software, or a combination thereof. Thus, the methods and apparatus of the present invention, or certain aspects or portions thereof, can take the form of program code (i.e., instructions) embedded in a tangible medium, such as a removable hard disk, USB flash drive, floppy disk, CD-ROM, or any other machine-readable storage medium, wherein when the program is loaded into and executed by a machine such as a computer, the machine becomes an apparatus for practicing the present invention.

[0113] When the program code is executed on a programmable computer, the computing device generally includes a processor, a processor-readable storage medium (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device. The memory is configured to store program code; the processor is configured to execute the method of the present invention according to instructions in the program code stored in the memory.

[0114] By way of example, and not limitation, readable media include readable storage media and communication media. Readable storage media stores information such as computer-readable instructions, data structures, program modules, or other data. Communication media generally embodies computer-readable instructions, data structures, program modules, or other data in the form of modulated data signals such as carrier waves or other transmission mechanisms, and includes any information delivery medium. Any combination of the above is also included within the scope of readable media.

[0115] In the specification provided herein, the algorithms and displays are not inherently related to any particular computer, virtual system, or other device. Various general-purpose systems can also be used with the examples of this invention. The required structure for constructing such systems is apparent from the above description. Furthermore, this invention is not directed to any particular programming language. It should be understood that the contents of the invention described herein can be implemented using various programming languages, and the above description of specific languages ​​is for the purpose of disclosing preferred embodiments of the invention.

[0116] Numerous specific details are set forth in the specification provided herein. However, it will be understood that embodiments of the invention may be practiced without these specific details. In some instances, well-known methods, structures, and techniques have not been shown in detail so as not to obscure the understanding of this specification.

[0117] Those skilled in the art will understand that modules, units, or components of the devices disclosed in the examples herein can be arranged in devices as described in this embodiment, or alternatively, can be located in one or more devices different from the device in this example. The modules in the foregoing examples can be combined into a single module or further divided into multiple sub-modules.

[0118] Furthermore, some of the embodiments described herein are methods or combinations of method elements that can be implemented by a processor of a computer system or by other means of performing the functions. Therefore, a processor having the necessary instructions for implementing the methods or method elements forms means for implementing the methods or method elements. Furthermore, the elements described herein in the apparatus embodiments are examples of means for implementing the functions performed by elements for the purposes of carrying out the invention.

[0119] As used herein, unless otherwise specified, the use of ordinal numbers such as “first,” “second,” “third,” etc., to describe ordinary objects merely indicates different instances of similar objects and is not intended to imply that the objects being described must have a given order in time, space, ordering, or any other manner.

[0120] Although the invention has been described with reference to a limited number of embodiments, those skilled in the art will understand from the foregoing description that other embodiments are conceivable within the scope of the invention described herein. Furthermore, it should be noted that the language used in this specification has been chosen primarily for readability and instructional purposes, and not for the purpose of interpreting or limiting the subject matter of the invention. Therefore, many modifications and alterations will be apparent to those skilled in the art without departing from the scope and spirit of the appended claims.

Claims

1. A method for determining unmanned aerial vehicle (UAV) logistics delivery routes, executed in a computing device, the method comprising: Obtain a delivery task list, which records delivery information for at least one delivery task, including delivery address, delivery weight, and latest delivery time. Arrange the delivery addresses sequentially to generate multiple initial delivery routes. Construct a route optimization objective function with the optimization objectives of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window. Minimizing total energy consumption is calculated based on the distance between delivery addresses and the delivery weight. The time window constraint is achieved by determining whether the expected arrival time at the target delivery address is earlier than its latest delivery time. Based on the route optimization objective function, the initial delivery routes are optimized and screened to obtain optimized candidate routes; using the A* algorithm, obstacle avoidance and smoothing processing are performed on each candidate route in three-dimensional space to output the final delivery route.

2. The method as described in claim 1, wherein, Also includes: A set of nodes is constructed based on each delivery address, and an edge set is constructed between the elements in the node set. The node set and the edge set together define a directed graph structure. The distance and compliance coefficient between the delivery addresses corresponding to each edge are determined. The compliance coefficient indicates whether the path between the delivery addresses crosses a no-fly zone. The energy consumption and the flight time required by the drone are determined based on the distance, delivery weight, wind speed parameters, and required altitude. The distance, compliance coefficient, energy consumption, and flight time are used as edge attributes of the corresponding edges and associated with the directed graph structure.

3. The method as described in claim 2, wherein, The delivery addresses are arranged sequentially to generate multiple initial delivery routes, including: permuting all node numbers in the directed graph structure to generate multiple number sequences; inserting the starting point number at the beginning of each number sequence and the return point number at the end, thereby obtaining each initial delivery route.

4. The method of claim 2, wherein, The energy consumption is expressed by the following formula: ;in, For delivery address Energy consumption in the room For delivery address The distance between, For the delivery address The delivery weight at the time of takeoff. For delivery address The headwind speed of the drone For delivery address The drone climbed to a higher altitude. This represents the weight value.

5. The method of claim 4, wherein, The objective function for route optimization is expressed by the following formula: ;in, The value represents the fitness of route optimization. For delivery address Energy consumption in the room To reach the delivery address Time, For delivery address The latest delivery time, This is a penalty for late completion; R represents the task completion rate: R = number of completed tasks / total number of tasks. This represents the weight value.

6. The method of claim 5, wherein, Based on the route optimization objective function, the initial delivery routes are optimized and screened to obtain optimized candidate routes, including: for each initial delivery route, using the route optimization objective function, calculating the fitness among the delivery addresses it contains, and determining the initial delivery route as a feasible delivery route when the fitness meets the threshold requirement; using a genetic algorithm to encode each feasible delivery route as a chromosome, and performing population evolution through selection, crossover, and mutation operators to obtain each candidate route.

7. The method according to any one of claims 1-6, wherein, Using the A* algorithm, obstacle avoidance and smoothing are performed on each candidate route in three-dimensional space to output the final delivery route. This includes: constructing a three-dimensional spatial mesh model covering each delivery address, where the mesh model contains obstacle areas; determining the cost function of the A* algorithm based on the straight-line distance between delivery addresses, the cost per unit of wind resistance, the height difference between the UAV and building rooftops during flight, and the penalty factor for approaching no-fly zones, resulting in an improved A* algorithm; using the improved A* algorithm, combined with the candidate routes, performing route search in the three-dimensional spatial mesh model to generate a three-dimensional delivery route; and smoothing the three-dimensional delivery route using Bezier or B-spline curves to obtain the final delivery route.

8. The method of claim 7, wherein, The cost function is expressed by the following formula: ;in, Let cost function be The straight-line distance between delivery addresses. Cost per unit of wind resistance The height difference between the drone and the building roof during flight. A penalty factor for drones approaching no-fly zones during flight. ， ， ， All are weight values.

9. The method of claim 2, wherein, It also includes the steps of: determining whether the delivery point corresponding to each delivery address meets the delivery conditions; replacing the delivery address that does not meet the delivery conditions with the nearest delivery address, and updating the directed graph structure.

10. The method of claim 9, wherein, The delivery conditions include at least one of the following: whether the delivery point is 2*2m 2 The above-mentioned unobstructed horizontal platform; whether the tilt angle of the landing area of ​​the delivery point is less than 10°; whether the number of people, heat sources and electromagnetic interference within the landing radius of the delivery point exceeds the standard.

11. A drone logistics delivery route determination device, residing in a computing device, the device comprising: The system includes: an acquisition module, adapted to acquire a list of delivery tasks, which records delivery information for at least one delivery task, including delivery address, delivery weight, and latest delivery time; a sorting module, adapted to sort the delivery addresses sequentially to generate multiple initial delivery routes; a construction module, adapted to construct a route optimization objective function with the optimization goals of minimizing total energy consumption, maximizing task completion rate, and satisfying the latest delivery time window, wherein minimizing total energy consumption is calculated based on the distance between delivery addresses and the delivery weight, and the time window constraint is achieved by determining whether the expected time to reach the target delivery address is earlier than its latest delivery time; and a filtering module, adapted to optimize and filter the initial delivery routes based on the route optimization objective function to obtain optimized candidate routes. The processing module is suitable for using the A* algorithm to perform obstacle avoidance and smoothing processing on each candidate route in three-dimensional space, thereby outputting the final delivery route.

12. A computing device, comprising: At least one processor; and a memory storing program instructions, wherein the program instructions are configured to be executed by the at least one processor, the program instructions including instructions for performing the method as described in any one of claims 1-10.

13. A readable storage medium storing program instructions that, when read and executed by a computing device, cause the computing device to perform the method as described in any one of claims 1-10.

Images