A 3D path planning method for a drone used for patrolling fig seedlings

Abstract

This invention provides a three-dimensional path planning method for a drone used for fig seedling patrol, relating to the field of agricultural production and operation management technology. The method includes: acquiring planning parameters; establishing a three-dimensional path planning model for the patrol drone based on the planning parameters; the three-dimensional path planning model being a problem model with the objective of finding the set of drone service paths that minimizes the total cost of the patrol management system; and solving the three-dimensional path planning model using a variable neighborhood search algorithm with a utilization rate evaluation mechanism to obtain the final set of drone service paths. The utilization rate evaluation mechanism is used to calculate the utilization rate index of each candidate drone service path generated by the variable neighborhood search algorithm, and to filter and guide neighborhood movement operations for the generated candidate drone service paths based on the utilization rate index. This improves the feasibility, safety, and economic benefits of patrol management in fig planting environments.

Description

Technical Field

[0001] This invention relates to the field of agricultural production and operation management technology, specifically to a three-dimensional path planning method for a drone used for patrolling fig seedlings. Background Technology

[0002] In precision agriculture production and operation management, plant protection management is a crucial link in ensuring crop yield and quality, and drone patrol operations have become an important technical means due to their high efficiency. With the development of precision agriculture, the application scenarios of drone patrols have gradually expanded from plains and fields to high-value horticultural crop planting areas, especially orchards, represented by fig trees. Fig orchards are characterized by contiguous planting, high elevation, and dense foliage, which poses new requirements for drone patrol operations. Against this backdrop, how to combine the needs of fig orchards with the service capabilities of drones, and optimize the number of spraying passes, sequence, and dosage of pesticides through three-dimensional path planning technology to achieve efficient and low-cost patrol management has become an urgent technical problem to be solved.

[0003] To address the aforementioned issues, existing technologies have introduced intelligent scheduling systems based on the "vehicle routing problem" to optimize the operational routes of patrol drones. This technology primarily studies how to plan optimal routes for patrol drones under constraints to minimize service costs or maximize efficiency. In agricultural patrol management, such route planning methods have been applied to scenarios such as large-scale integrated pest management and multi-drone field operations, achieving efficient management in plain field environments by shortening operation time, improving drone utilization, and reducing patrol costs.

[0004] However, existing patrol route planning methods have significant limitations when directly applied to fig cultivation environments. Current methods typically treat flight power as a fixed value when modeling the power consumption of patrol drones, neglecting the impact of significant changes in terrain elevation within the orchard on the drone's power during ascent and descent. Furthermore, dynamic changes in pesticide load during the service process directly affect real-time flight power. These factors make it difficult for existing technologies to accurately estimate the power consumption of patrol drones along elevation-changing sections, thus affecting the assessment of available energy consumption. This makes it difficult to safely and effectively implement existing route planning schemes in fig cultivation environments, limiting the feasibility and economic benefits of patrol drone management. Summary of the Invention

[0005] To address the aforementioned problems in the existing technology, this invention provides a three-dimensional path planning method for a fig seedling patrol drone.

[0006] The technical problem to be solved by this invention is achieved through the following technical solution: This invention provides a three-dimensional path planning method for a fig seedling patrol drone, comprising: Obtain planning parameters; the planning parameters should include at least: the three-dimensional location and pesticide application requirements of each fig farm to be served, as well as the performance and cost parameters of the patrol drone; A three-dimensional path planning model for patrol drones is established based on planning parameters. The three-dimensional path planning model is a problem model with the goal of finding the set of drone service paths that minimize the total cost of the patrol management system. The total cost includes the dynamic electricity cost calculated based on the total energy consumption of the patrol drones and the fixed cost of calling up the patrol drones. The three-dimensional path planning model decomposes the flight process of the patrol drones into uniform horizontal flight state, operation state, and vertical take-off and landing state by different altitudes and dynamic liquid loads. The total energy consumption of the patrol drones is constructed based on the cumulative work done in the uniform horizontal flight state, operation state, and vertical take-off and landing state, respectively. A variable neighborhood search algorithm with a utilization evaluation mechanism is used to solve the three-dimensional path planning model to obtain the final set of UAV service paths. The utilization evaluation mechanism is used to calculate the utilization index of each candidate UAV service path when the variable neighborhood search algorithm generates each candidate UAV service path, and to filter and guide the neighborhood movement operations of the generated candidate UAV service paths based on the utilization index to obtain the final set of UAV service paths. The utilization index is composed of both capacity utilization and power utilization.

[0007] This invention provides a three-dimensional path planning method for a fig seedling patrol drone, comprising: acquiring planning parameters; the planning parameters include at least: the three-dimensional location and pesticide application requirements of each fig farm to be served, as well as the performance and cost parameters of the patrol drone; establishing a three-dimensional path planning model corresponding to the patrol drone based on the planning parameters; the three-dimensional path planning model is a problem model with the goal of finding the set of drone service paths that minimize the total cost of the patrol management system, the total cost including the dynamic electricity cost calculated based on the total energy consumption of the patrol drone and the fixed cost of calling the patrol drone; wherein, the three-dimensional path planning model decomposes the flight process of the patrol drone into different altitudes and dynamic pesticide application loads. The invention decomposes the flight process into uniform horizontal flight, operational, and vertical take-off / landing states, and calculates the total energy consumption of the patrol drone based on the cumulative work done in each of these states. A variable neighborhood search algorithm with a utilization rate evaluation mechanism is used to solve the three-dimensional path planning model, resulting in a final set of drone service paths. The utilization rate evaluation mechanism calculates the utilization rate index of each candidate drone service path generated by the variable neighborhood search algorithm, and uses this index to filter and guide neighborhood movement operations for generating candidate drone service paths, thus obtaining the final set of drone service paths. The utilization rate index is composed of both capacity utilization and power utilization. In this invention, by first decomposing the flight process into uniform horizontal flight, operational, and vertical take-off / landing states, and calculating the cumulative work done based on different altitudes and dynamic liquid loads, the problem of existing methods neglecting the impact of terrain changes and load on power is solved, thereby achieving accurate estimation of the patrol drone's energy consumption. Then, by establishing a three-dimensional path planning model with the goal of minimizing total cost, dynamic electricity costs are combined with fixed costs for optimization. This solves the problem of existing technologies struggling to determine available energy consumption due to inaccurate energy consumption estimation, making path planning more aligned with the safety and economic needs of the fig planting environment. Finally, a variable neighborhood search algorithm with a utilization rate evaluation mechanism is employed to filter and guide neighborhood movement operations based on capacity utilization and power utilization rates. This addresses the issues of low feasibility and poor efficiency of the original planning in the fig planting environment, thereby generating a set of efficient and reliable UAV service paths. In summary, this invention achieves dynamic modeling and cost-optimized path planning for the energy consumption of patrol UAVs, improving the feasibility, safety, and economic benefits of patrol management in the fig planting environment.

[0008] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0009] Figure 1 A flowchart illustrating a three-dimensional path planning method for a fig seedling patrol drone provided in an embodiment of the present invention; Figure 2 An exemplary schematic diagram of a three-dimensional path planning model constrained by service capacity and multiple influencing factors is shown; Figure 3 An example diagram illustrates the execution of a three-dimensional path planning method for a fig seedling patrol drone. Detailed Implementation

[0010] The technical problem this invention aims to solve is the design of a three-dimensional path planning method for patrol drones in precision agriculture. This three-dimensional path planning method employs operations research optimization modeling, simultaneously considering the service capacity constraints of the patrol drone and the demand of the fig farms to be served during the optimization process. It allocates limited pesticide solution to different fig farms to be served via different paths and in different orders, and designs a variable neighborhood algorithm with a utilization rate evaluation mechanism to solve the model.

[0011] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0012] To achieve dynamic modeling and cost-optimized path planning for the energy consumption of patrol drones, and to improve the feasibility, safety, and economic benefits of patrol management in fig planting environments, this invention provides a three-dimensional path planning method for fig seedling patrol drones.

[0013] Before introducing the three-dimensional path planning method of this invention, a brief explanation of the symbol parameters involved in this invention will be given: A collection of patrol drones; : Indicates the total number of patrol drones; Patrol drones ; A collection of base stations In this invention, there is a fixed base station; A collection of fig farms awaiting service; Fig farmland awaiting service , ; Fig farmland awaiting service , ; Site , ; Site , ; Fig farmland awaiting service From the starting point to the fig farms awaiting service The horizontal distance between the landing points ; Fig farmland awaiting service From the starting point to the fig farms awaiting service The vertical distance from the landing point, ; Fig farmland awaiting service Interval of medium-sized fruit trees, ; Fig farmland awaiting service The horizontal distance between the starting point and the landing point, ; Fig farmland awaiting service The altitude difference between the starting point and the landing point, ; Fig farmland awaiting service The length of the rectangle, ; Fig farmland awaiting service The width of the rectangle, ; Fig farmland awaiting service The number of rows of medium-sized fruit trees, ; Fig farmland awaiting service The number of fruit trees in each row. ; Fig farmland awaiting service Required volume of medicine solution Unit: L; : Weight of the patrol drone, in kg; : Maximum payload of the medical kit for patrol drones, in liters (L); Patrol drones and the total weight of the drug solution it carries. , Unit: kg; : Average horizontal flight speed of the patrol drone, in m / s; : Average horizontal flight speed of patrol drones during spraying operations, in m / s; : Spraying speed of pesticides during patrol operations by patrol drones, in L / s; Vertical induced velocity of patrol drone, in m / s; : Average vertical takeoff and landing speed of patrol drones, in m / s; : Unit power consumption cost of patrol drones, in CNY / kW·h; : Fixed cost per drone, in CNY; : Horizontal flight power of patrol drones when moving between fig fields awaiting service, in KW; Patrol drones Leaving the station The horizontal flight power afterward; Flight power of patrol drones during spraying operations, in KW; : Maximum power of the patrol drone, in kW·h; Patrol drones by Duration of flight power operation, Unit: s; The cumulative work done by the patrol drone in a constant-speed horizontal flight state. Unit: J; Patrol drones by Duration of flight power operation, Unit: s; The cumulative work done by the patrol drone during operation. Unit: J; The cumulative work done by the patrol drone during vertical take-off and landing. Unit: J; Patrol drones The work done against resistance during upward and downward movement, measured in J; If patrol drones From fig farms awaiting service Fly to fig farms awaiting service ,but ;otherwise, ; If patrol drones From the site Fly to the station ,but ;otherwise, ; If patrol drones Served fig farms awaiting service but ;otherwise .

[0014] Figure 1 This is a flowchart illustrating a three-dimensional path planning method for a fig seedling patrol drone provided in an embodiment of the present invention, as shown below. Figure 1 As shown, it includes: S101. Obtain planning parameters.

[0015] The planning parameters should include at least: the three-dimensional location and pesticide application requirements of each fig farm to be served, as well as the performance and cost parameters of the patrol drone.

[0016] Optionally, the performance parameters of the patrol drone include: maximum payload, weight information, maximum takeoff weight, maximum battery capacity, battery mass, horizontal cruising speed, maximum climb rate, maximum descent rate, hovering power, operating power, rotor system efficiency, positioning accuracy, spraying width, and liquid flow rate range. The cost parameters for patrol drones include: fixed cost per drone, number of drones that can be deployed, unit energy consumption electricity price, and cost of medicine.

[0017] The 3D path planning problem for fig orchard patrol drones in precision agriculture can be represented as follows: Within a service area, several fig orchards require patrol drone operations. Several drones depart from a fixed base station. Each drone can serve several fig orchards during its flight, and each orchard can only be served once. The locations and needs of the fig orchards are known. The fig trees in the orchards are arranged in rows, with the altitude of each row increasing sequentially. The decision-making process involves allocating service paths and sequences to each fig orchard within the limited service capacity of each drone, maximizing the economic benefits of the patrol management system. Figure 2 An exemplary schematic diagram of a three-dimensional path planning model constrained by service capacity and multiple influencing factors is shown. Figure 2As shown, the top of the 3D path planning model clearly defines the physical resource constraints faced by the patrol drone during mission execution, such as the maximum available power and maximum liquid payload. Under the combined influence of these constraints, altitude characteristics, load variations, and other factors, specific patrol operation requirements and economic considerations arise. Based on this, Figure 2 The bottom of the document shows the specific implementation plan for the path planning, in which multiple patrol drones depart from the base station and proceed to different farmlands or fig fields (such as farmlands A, B, C, D, etc.) in sequence according to the planned path to perform spraying tasks, and finally return to the base station, thus forming a complete closed-loop service path.

[0018] S102. Establish a three-dimensional path planning model for the patrol drone based on the planning parameters.

[0019] The three-dimensional path planning model is a problem model that aims to solve for the set of UAV service paths that minimize the total cost of the patrol management system. The total cost includes the dynamic electricity cost calculated based on the total energy consumption of the patrol UAVs and the fixed cost of calling up the patrol UAVs. The three-dimensional path planning model decomposes the flight process of the patrol UAVs into uniform horizontal flight, operation, and vertical take-off and landing states by using different altitudes and dynamic liquid loads. Based on the cumulative work done in the uniform horizontal flight, operation, and vertical take-off and landing states, the total energy consumption of the patrol UAVs is constructed.

[0020] Optionally, the uniform horizontal flight state refers to the flight process of the patrol drone when it is moving between fig fields to be served; The operational status refers to the flight process of the patrol drone during spraying operations; Vertical ascent and descent refers to the flight process of patrol drones when they move between different altitudes.

[0021] Alternatively, the three-dimensional path planning model can be represented as: (1-1) This represents the total cost of the patrol management system. This indicates the unit power consumption cost of the patrol drone. Indicates site , , Indicates site , , This represents a collection of fig farms awaiting service. Represents the set of base stations. Indicates patrol drone , , This indicates a collection of patrol drones. Indicates patrol drone From the site Fly to the station The corresponding binary variable, This represents the cumulative work done by the patrol drone during its uniform horizontal flight. This indicates the cumulative work done by the patrol drone during operation. This indicates the cumulative work done by the patrol drone during its vertical take-off and landing. This indicates the fixed cost of each patrol drone. Indicates fig farmland awaiting service , , Indicates fig farmland awaiting service , , Indicates patrol drone From fig farms awaiting service Fly to fig farms awaiting service The corresponding binary variable, (1-2) (1-3) (1-4) (1-5) (1-6) (1-7) (1-8) (1-9) in, This indicates the total number of patrol drones. This represents a collection of fig farms awaiting service. any non-empty subset, ,and ; Indicates site , , Indicates site , , Indicates patrol drone From the site Fly to the station The corresponding binary variable, Indicates when patrol drones Served fig farms awaiting service but ;otherwise ; Indicates fig farmland awaiting service Required volume of medicine solution This indicates the maximum payload capacity of the patrol drone. This indicates the maximum power of the patrol drone. Indicates fig farmland awaiting service From the starting point to the fig farms awaiting service The horizontal distance between the landing points This represents the average speed of the patrol drone in the horizontal direction. Indicates fig farmland awaiting service The horizontal distance between the starting point and the landing point, Indicates patrol drone by Duration of flight power operation, Indicates patrol drone by Duration of flight power operation, This indicates the horizontal flight power of the patrol drone when it is being moved between the fig fields being served. This indicates the flight power of the patrol drone during spraying operations. Indicates fig farmland awaiting service The number of rows of medium-sized fruit trees, Indicates fig farmland awaiting service The length of the rectangle, This represents the average speed of the patrol drone during its spraying operation in the horizontal direction. Indicates patrol drone Fly from the station Arrive at the station The corresponding binary variable, It means arbitrary.

[0022] It should be noted that the above formula (1-1) is the objective of the three-dimensional path planning model, namely, to minimize the total cost of the patrol management system.

[0023] Constraints (1-2) describe the limits on the number of patrol drones that can be dispatched.

[0024] Constraint (1-3) describes that the number of incoming and outgoing edges is equal for each station.

[0025] Constraints (1-4) describe that each site has one and only one patrol drone service.

[0026] Constraints (1-5) achieve sub-loop elimination.

[0027] Constraints (1-6) describe that the sum of the required amount of medicine for each patrol drone service order is less than the capacity of the medicine box.

[0028] Constraints (1-7) describe that the total cumulative power consumption of each patrol drone in the three flight states shall not exceed the battery power limit.

[0029] Constraints (1-8) describe the patrol drone by Duration of flight power operation.

[0030] Constraints (1-9) describe the patrol drone by Duration of flight power operation.

[0031] The calculation process for the cumulative work done by the patrol drone in uniform horizontal flight is as follows: ; ; in, This indicates the horizontal flight power of the patrol drone when it is being moved between the fig fields being served. To monitor the average horizontal flight speed of the patrol drone, To guide the vertical speed of the patrol drone, To induce power factor, air density, To protect the rotor disk area of ​​the patrol drone, To maintain the rotational speed of the drone rotor, This indicates the rotor diameter of the patrol drone. To improve the drag ratio of patrol drones, To improve the rotor blade area ratio of patrol drones, Indicates the tip speed ratio. Indicates the incoming wind speed.

[0032] Induced power factor The solution is based on a simplified model proposed by Bangura, M. and Mahony, R. This simplified model indicates the thrust coefficient of a single rotor. With induced power factor Under specific operating conditions, it exhibits an approximately linear relationship. Therefore, in obtaining... After obtaining the measured data, the slope can be determined through data fitting. and intercept The value of this value is used to deduce the induced power factor. .

[0033] Vertical induced velocity of patrol drones The method is as follows: Considering the model proposed by Lei, Y. et al., when there is no horizontal wind, i.e. , At this moment, the patrol drone hovered with thrust. Balanced with gravity, therefore: ; This indicates the total weight of the current patrol drone and its payload of medicine. Represents gravitational acceleration. ; ; Combining the above three formulas, we get ; In addition to the two variables mentioned above, All other variables involved have reference values.

[0034] because The patrol drone flies at a constant speed, therefore It is a relatively constant value (which will depend on) (If the condition is constant in stages), then any patrol drone can be obtained. The cumulative work done in uniform horizontal flight .

[0035] The cumulative work done by the patrol drone in uniform horizontal flight is expressed as: ; This represents the cumulative work done by the patrol drone during its uniform horizontal flight. Indicates patrol drone Leaving the station The subsequent horizontal flight power, Indicates patrol drone by Duration of flight power operation, This indicates a collection of patrol drones. Indicates when patrol drones Served fig farms awaiting service but ;otherwise .

[0036] Cumulative work done by patrol drones during operation The calculation process is as follows: During the patrol drone operation, the liquid pesticide was sprayed at a certain speed. Spraying at a constant speed, and maintaining the same travel speed. , The amount of liquid medicine will decrease dynamically in real time due to changes in the quality of the solution, therefore the following relationship exists: ; ; ; ; This indicates the total weight of the patrol drone and its payload of medicine. Follow A changing function, This indicates the total weight of the patrol drone and its payload of medicine. Indicates the flight power of the patrol drone during spraying operations. Over time A changing function; This indicates the total mass of the patrol drone over time. A changing function, In summary, the cumulative work done by the patrol drone during operation can be obtained as follows: ; This indicates the cumulative work done by the patrol drone during operation. Indicates the induced power factor. This indicates the rotational speed of the patrol drone's rotor. This indicates the rotor diameter of the patrol drone. Indicates air density, This indicates the rotor disk area of ​​the patrol drone. Represents gravitational acceleration. This indicates the power of the patrol drone in overcoming airframe drag. Indicates patrol drone by Duration of flight power operation, This indicates the average speed of the patrol drone during its spraying operation in the horizontal direction.

[0037] The calculation process for the cumulative work done by the patrol drone during vertical take-off and landing is as follows: This section adopts an energy conservation framework, using the vertical energy consumption of the patrol drone throughout its flight path as the metric. For ease of simulation, the vertical acceleration is ignored, and the drone's ascent and descent are treated as uniform motion. Meanwhile, frictional heat energy is ignored.

[0038] Under these conditions, assuming the patrol drone undergoes vertical takeoff and landing, the energy output by the motor will be converted into the sum of three parts: the change in the drone's kinetic energy, the change in its gravitational potential energy, and the work done by the drone against air resistance. Therefore, the cumulative work done by the patrol drone during vertical takeoff and landing can be expressed as: ; This indicates the cumulative work done by the patrol drone during its vertical take-off and landing. This represents the change in kinetic energy of the patrol drone. This represents the change in gravitational potential energy of the patrol drone. Indicates patrol drone When moving upwards or downwards, it does work by overcoming resistance.

[0039] Throughout the entire flight path, the initial and final speeds are both 0, meaning... ; The change in gravitational potential energy of the patrol drone depends solely on the pesticide application requirements and altitude of each fig farm to be served, and is unrelated to the drone's flight path or service sequence. ; Indicates fig farmland awaiting service The altitude difference between the starting point and the landing point, Indicates that if patrol drones Served fig farms awaiting service but ;otherwise , Indicates fig farmland awaiting service Required volume of medicine solution It represents the acceleration due to gravity.

[0040] Air resistance always does work in the opposite direction of motion, and the distance of work done is the change in elevation coordinates between the fig fields to be served, i.e.: ; This indicates the average vertical takeoff and landing speed of the patrol drone. Indicates fig farmland awaiting service From the starting point to the fig farms awaiting service The vertical distance from the landing point, To improve the drag ratio of patrol drones, To improve the rotor blade area ratio of patrol drones, air density, The area of ​​the rotor disk of the patrol drone.

[0041] S103. A variable neighborhood search algorithm with utilization evaluation mechanism is used to solve the three-dimensional path planning model to obtain the final set of UAV service paths.

[0042] The utilization rate evaluation mechanism is used to calculate the utilization rate index of each candidate UAV service path when the variable neighborhood search algorithm generates each candidate UAV service path, and to filter and guide the neighborhood movement operations generated by the candidate UAV service path based on the utilization rate index to obtain the final set of UAV service paths; the utilization rate index is composed of both capacity utilization and power utilization.

[0043] This invention provides a three-dimensional path planning method for a UAV used for fig seedling patrol. First, by breaking down the flight process into uniform horizontal flight, operational phases, and vertical ascent / descent, and calculating the cumulative work done based on different altitudes and dynamic pesticide loads, it solves the problem of existing methods neglecting the impact of terrain changes and load on power, thus achieving accurate estimation of the patrol UAV's energy consumption. Second, by establishing a three-dimensional path planning model with the goal of minimizing total cost, it optimizes by combining dynamic electricity costs with fixed costs, solving the problem of existing technologies struggling to determine available energy consumption due to inaccurate energy consumption estimations. This makes the path planning more aligned with the safety and economic needs of the fig planting environment. Finally, a variable neighborhood search algorithm with a utilization rate evaluation mechanism is employed to filter and guide neighborhood movement operations based on capacity utilization and power utilization rates, solving the problems of low feasibility and poor efficiency of the original planning in the fig planting environment, thereby generating a set of efficient and reliable UAV service paths. In summary, this invention achieves dynamic modeling and cost-optimized path planning for the energy consumption of patrol UAVs, improving the feasibility, safety, and economic benefits of patrol management in fig planting environments.

[0044] Optionally, S103 includes: S1031. Based on the planning parameters, generate an initial set of UAV service paths as the current solution, and record the current solution as the current optimal solution; S1032. Guided by the utilization evaluation mechanism, a variable neighborhood search algorithm is used to perform variable neighborhood search processing on the current solution to generate candidate solutions; S1033. If the objective function value of a candidate solution is better than the objective function value of the current optimal solution, then update the current solution and the current optimal solution as candidate solutions; wherein, the objective function value of the candidate solution is the function value corresponding to the substitution of the candidate solution into the three-dimensional path planning model; the objective function value of the current optimal solution is the function value corresponding to the substitution of the current optimal solution into the three-dimensional path planning model. S1034. Repeat steps S1032 and S1033 until the number of consecutive iterations without improvement reaches the perturbation initiation threshold. S1035. When S1034 is true, perform a perturbation operation on the current optimal solution and set the perturbed solution obtained after the perturbation operation as the current solution; S1036. Repeat steps S1032 to S1035 until the preset termination condition is met, and take the current optimal solution corresponding to the preset termination condition as the final UAV service path set.

[0045] Optionally, the preset termination conditions are: the total running time of the solution process reaches a preset duration threshold, the number of iterations reaches an iteration number threshold, or the number of iterations in which the objective function value of the current optimal solution has not been improved continuously reaches a preset stagnation threshold.

[0046] Alternatively, the utilization rate metric can be expressed as: ; in, This represents the value of the utilization rate indicator. Indicates the weighting coefficient. Indicates capacity utilization rate. Indicates power utilization rate. This indicates the ideal utilization rate of the medicine box capacity and the drone's battery power. ; ; Indicates fig farmland awaiting service , , Indicates fig farmland awaiting service , , This represents a collection of fig farms awaiting service. Indicates patrol drone From fig farms awaiting service Fly to fig farms awaiting service The corresponding binary variable, Indicates fig farmland awaiting service Required volume of medicine solution This indicates the maximum payload capacity of the patrol drone. This indicates the total number of patrol drones. Indicates patrol drone The estimated cumulative work done to complete the flight path. This indicates the maximum power of the patrol drone. Indicates patrol drone , , This indicates a collection of patrol drones. It means arbitrary.

[0047] To fully illustrate the three-dimensional path planning method provided in the embodiments of the present invention, please refer to... Figure 3The execution process of the three-dimensional path planning method of the present invention will be described in detail. Figure 3 An example diagram illustrates the execution of a three-dimensional path planning method for a fig seedling patrol drone.

[0048] First, the parameters involved are defined: This represents a collection of fig farms awaiting service. any non-empty subset, ,and ; This represents the total cost of the patrol management system. Indicates the current neighborhood search operator index. Represents the number of neighborhood search operators. Indicates the current iteration number. Indicates the limit on the number of iterations. Indicates the current optimal solution The cumulative number of times remains unchanged. represent The limit on the number of times remains unchanged. Let , , .

[0049] Before launching the variable neighborhood search algorithm, it is necessary to define how the solution to the problem is represented. In this invention, a complete solution is represented by an integer sequence. This integer sequence represents a set of patrol drone flight paths serving all the fig farms to be served. Specifically, the integer 1 represents the patrol drone base station, and other positive integers represent the numbers of each fig farm to be served. A complete solution structure consists of these integers, where the base station number 1 serves not only as the start and end point of each sub-path but also as a separator between different sub-paths. For example, the solution sequence structure (1, 5, 3, 1, 2, 4, 1) represents a scheme containing two sub-paths: the first is 1-5-3-1, and the second is 1-2-4-1. Furthermore, a sub-path is a path in the solution, representing a patrol drone. The path taken.

[0050] Step 1: Set up the energy consumption calculation function.

[0051] To accurately assess the flight energy consumption of patrol drones in a fig orchard environment, this invention establishes a phased energy consumption calculation model. This model decomposes the complete process of a single patrol drone flight into three independent states, with the total energy consumption of a single sub-path being... .

[0052] Step 1.1: By traversing each flight segment in the sub-path, calculate the instantaneous power of the patrol drone during its uniform horizontal flight based on the total dynamic flight mass of that segment. Then, combine this with the flight time of that segment to calculate the energy consumption of that segment, and sum them up to obtain the final energy consumption. .

[0053] Step 1.2: First, calculate the patrol operation time of the sub-path patrol drone. Based on the continuous change of the total mass of the patrol drone over time, integrate the instantaneous power of the total mass of the patrol drone over the operation time to obtain... .

[0054] Step 1.3: Through cumulative calculation, the energy output by the motor is converted into the sum of three parts: the change in kinetic energy of the patrol drone, the change in gravitational potential energy of the patrol drone, and the work done by the patrol drone against air resistance. Finally, the cumulative work done by the patrol drone in its vertical take-off and landing state is obtained. .

[0055] Step 1.4: Calculate the values ​​from steps 1.1, 1.2, and 1.3 respectively. , as well as Add them together to get the total energy consumption of this sub-path. .

[0056] Step 2: Generate an initial feasible solution .

[0057] The goal of this step is to generate an initial feasible solution that satisfies all constraints. The initial feasible solution is also called the initial set of UAV service paths.

[0058] Step 2.1: Extract all fig farm numbers to be served and sort them randomly to obtain a unique farm visit sequence.

[0059] Step 2.2: Sequentially add the pending fig farm numbers from the farm visit sequence to the current subpath, and accumulate the required pesticide solution for that subpath. If a pending fig farm is added... If so, the fig farm to be served will be the first farm served in the next sub-path; Indicates patrol drone Fig farms awaiting service The required amount of medicine solution at that time This indicates the maximum payload of the patrol drone's medicine tank. If the current sub-path continues to add subsequent fig farms to be served, then the process continues until all fig farms to be served in the farm visit sequence have been assigned, resulting in an initial set of UAV service paths consisting of several sub-paths.

[0060] Step 2.3: For each sub-path in the initial drone service path set, call Step 1 to calculate its total energy consumption, and compare the calculation result with the maximum power constraint of the patrol drone. If all sub-paths meet the maximum power constraint, then this initial drone service path set is considered valid. If any sub-path does not meet the maximum power constraint, discard the initial set of drone service paths and return to step 2.1 to regenerate.

[0061] Step 2.4: Repeat steps 2.1 to 2.3 until an initial feasible solution is generated. and the initial feasible solution Assign to the current solution .

[0062] Step 3: Set up an evaluation mechanism.

[0063] To improve neighborhood search efficiency, a capacity utilization-based approach is introduced. and power utilization rate The evaluation mechanism evaluates a specific sub-path. This mechanism applies to neighborhood operators and performs an evaluation after each operator change.

[0064] Step 3.1: Calculate the total pesticide application requirement for all fig farms to be served in this sub-path, and divide it by the maximum payload of the patrol drone to obtain the capacity utilization rate of this sub-path. .

[0065] ; Step 3.2: Estimate the surrogate value of power consumption of the patrol drone during its flight along this sub-path, and divide the surrogate value of power consumption by the maximum available power of the patrol drone to obtain the power utilization rate of this sub-path. .

[0066] ; Step 3.3: Calculate the deviation of the two utilization rates from the preset ideal utilization rate and the degree of difference between the two utilization rates. Sum the above deviation and difference rates with weights to obtain the value of the sub-path utilization rate index. .

[0067] ; Step 4: Set up the neighborhood.

[0068] Neighborhoods include: inter-path exchange neighborhoods, inter-path insertion neighborhoods, intra-path exchange neighborhoods, and intra-path insertion neighborhoods.

[0069] Step 4.1: Exchange neighborhoods between paths.

[0070] The path exchange operator is used to exchange fig farmland to be served between two different sub-paths.

[0071] Step 4.1.1: Calculate the value of each sub-path in the initial feasible solution. And sorted in ascending order.

[0072] Step 4.1.2: From the results of Step 4.1.1, select the sub-path with the smallest sequence number as the target path. And select candidate swap paths sequentially from other sub-paths. .

[0073] Step 4.1.3: Calculated based on Step 3 , . Indicates the target path The value of the utilization rate index, Indicates candidate swap paths The value of the utilization rate indicator.

[0074] Step 4.1.4: Select a fig farm to be served from the candidate exchange paths. Select a fig farm to be served from the target path. and satisfy < Swap the two to generate the target path for the swap. Candidate switching paths , Indicates fig farmland awaiting service Required amount of medicine.

[0075] Step 4.1.5: Calculated based on Step 3 , . Indicates the target path of the exchange. The value of the utilization rate index, Candidate switching paths for exchange The value of the utilization rate indicator.

[0076] Step 4.1.6: For , Perform the following judgments in sequence: (1) Judgment , Are they respectively less than , (2) Determine whether the maximum liquid load constraint is met; (3) Determine whether the maximum available power constraint is met; (4) Calculate candidate solutions. Corresponding total cost Is it less than the total cost corresponding to the current solution? When all the above conditions are met, candidate solutions are set for the candidate swap paths. Otherwise, the current solution It remains unchanged.

[0077] Step 4.1.7: Repeat steps 4.1.3-4.1.6, when... After traversing all the remaining sub-paths, exit the neighborhood. .

[0078] Since inter-path neighbor insertion, intra-path neighbor swapping, and intra-path neighbor insertion all employ the conventional operational logic of the classic neighborhood search operator in the Vehicle Routing Problem (VRP) heuristic algorithm, their specific node selection details will not be elaborated upon here. Inter-path neighbor insertion is used to move fig farms to be served between two different sub-paths. The process involves calculating and sorting the utilization indices of each path to select a target path, then inserting the fig farms to be served in the target path into feasible positions on another candidate path according to specific rules, thereby generating new candidate solutions. Intra-path neighbor swapping is used to adjust the order of fig farms to be served within the same sub-path. The process involves calculating the path utilization indices to determine the target path, then selecting two fig farms to be served within that path and swapping their positions, thereby generating new candidate solutions. Intra-path neighbor insertion is used to adjust the order of fig farms to be served within the same sub-path. The process is as follows: the target path is determined by calculating the path utilization rate index, and then a fig farm that needs to be served is removed from the path and inserted into another location on the path, thereby generating a new candidate solution.

[0079] Finally, following the aforementioned judgment process, the following judgments are made sequentially for each candidate solution generated above: (1) whether the utilization rate index of the relevant path after the operation is better than before the operation; (2) whether the maximum liquid load constraint is met; (3) whether the maximum available power constraint is met; (4) whether the objective function value after the operation is less than the objective function value before the operation. When all the above conditions are met, the candidate solution is accepted and the current solution is updated; otherwise, the current solution remains unchanged.

[0080] Step 5: Neighborhood search.

[0081] Step 5.1: Calculation Corresponding total cost At the same time, let the current optimal solution , , This represents the total cost corresponding to the current optimal solution.

[0082] Step 5.2: For the current solution Perform a perturbation operation to obtain the perturbation solution. .

[0083] Step 5.3: with As input, the neighborhood operators from step 4 are called sequentially to perform a local search. When a neighborhood operator finds a better candidate solution... At that time, As a new starting point for the search, and starting the search again from the first neighborhood operator, the process is reset. If no better solution is found by a certain neighborhood operator, then the search jumps to the next neighborhood operator. .

[0084] Step 5.4: If Proceed to step 5.3; if Output candidate solutions .

[0085] Step 5.5: If Then let the current solution = , , Otherwise, they remain unchanged. This represents the total cost of the candidate solution.

[0086] Step 5.6: End the current neighborhood search. .

[0087] Step 5.7: If there are or End the neighborhood search process and output. The total cost is output as the result of the 3D path planning for the patrol drone. Otherwise, repeat steps 5.2-5.6 to obtain the final set of drone service paths.

[0088] The method provided in this embodiment of the invention can be applied to electronic devices. Specifically, the electronic device can be a desktop computer, a portable computer, a smart mobile terminal, a server, etc., and this embodiment of the invention does not limit the application to such devices.

[0089] Although the invention has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings and the disclosure, will understand and implement other variations of the disclosed embodiments in carrying out the claimed invention. In this description, the word "comprising" does not exclude other components or steps, "a" or "an" does not exclude a plurality, and "a plurality" means two or more, unless otherwise explicitly specified. Furthermore, while different embodiments may describe certain measures, this does not mean that these measures cannot be combined to produce good results.

[0090] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the inventive concept, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A three-dimensional path planning method for a UAV used for patrolling fig seedlings, characterized in that, include: Obtain planning parameters; The planning parameters include at least: the three-dimensional location and pesticide application requirements of each fig farm to be served, as well as the performance and cost parameters of the patrol drone. A three-dimensional path planning model is established based on the planning parameters for the patrol drone. The three-dimensional path planning model is a problem model that aims to solve the set of drone service paths that minimize the total cost of the patrol management system. The total cost includes the dynamic electricity cost calculated based on the total energy consumption of the patrol drone and the fixed cost of calling the patrol drone. The three-dimensional path planning model decomposes the flight process of the patrol drone into a uniform horizontal flight state, an operational state, and a vertical take-off and landing state by using different altitudes and dynamic liquid loads. The total energy consumption of the patrol drone is constructed based on the cumulative work done in the uniform horizontal flight state, the operational state, and the vertical take-off and landing state, respectively. A variable neighborhood search algorithm with a utilization evaluation mechanism is used to solve the three-dimensional path planning model to obtain the final set of UAV service paths. The utilization evaluation mechanism is used to calculate the utilization index of each candidate UAV service path generated by the variable neighborhood search algorithm, and to filter and guide neighborhood movement operations that generate the candidate UAV service paths based on the utilization index, thereby obtaining the final set of UAV service paths. The utilization index is composed of both capacity utilization and power utilization.

2. The three-dimensional path planning method for a fig seedling patrol drone according to claim 1, characterized in that, The performance parameters of the patrol drone include: maximum payload, weight information, maximum takeoff weight, maximum battery capacity, battery mass, horizontal cruising speed, maximum climb rate, maximum descent rate, hovering power, operating power, rotor system efficiency, positioning accuracy, spraying width, and liquid flow rate range. The cost parameters of the patrol drones include: fixed cost per drone, number of drones that can be deployed, unit energy consumption electricity price, and cost of medicine.

3. The three-dimensional path planning method for a fig seedling patrol drone according to claim 1, characterized in that, The uniform horizontal flight state refers to the flight process of the patrol drone when it is moving between fig farms to be served. The operational status refers to the flight process of the patrol drone during spraying operations. The vertical ascent and descent state refers to the flight process of the patrol drone when it moves between different altitudes.

4. The three-dimensional path planning method for a fig seedling patrol drone according to claim 1, characterized in that, The three-dimensional path planning model is represented as follows: ； This represents the total cost of the patrol management system. This indicates the unit power consumption cost of the patrol drone. Indicates site , , Indicates site , , This represents a collection of fig farms awaiting service. Represents the set of base stations. Indicates patrol drone , , This indicates a collection of patrol drones. Indicates patrol drone From the site Fly to the station The corresponding binary variable, This represents the cumulative work done by the patrol drone during its uniform horizontal flight. This indicates the cumulative work done by the patrol drone during operation. This indicates the cumulative work done by the patrol drone during its vertical take-off and landing. This indicates the fixed cost of each patrol drone. Indicates fig farmland awaiting service , , Indicates fig farmland awaiting service , , Indicates patrol drone From fig farms awaiting service Fly to fig farms awaiting service The corresponding binary variable, ； ； ； ； ； ； ； ； in, This indicates the total number of patrol drones. This represents a collection of fig farms awaiting service. any non-empty subset, ,and ; Indicates site , , Indicates site , , Indicates patrol drone From the site Fly to the station The corresponding binary variable, Indicates when patrol drones Served fig farms awaiting service but ;otherwise ; Indicates fig farmland awaiting service Required volume of medicine solution This indicates the maximum payload capacity of the patrol drone. This indicates the maximum power of the patrol drone. Indicates fig farmland awaiting service From the starting point to the fig farms awaiting service The horizontal distance between the landing points This represents the average speed of the patrol drone in the horizontal direction. Indicates fig farmland awaiting service The horizontal distance between the starting point and the landing point, Indicates patrol drone by Duration of flight power operation, Indicates patrol drone by Duration of flight power operation, This indicates the horizontal flight power of the patrol drone when it is being moved between the fig fields being served. This indicates the flight power of the patrol drone during spraying operations. Indicates fig farmland awaiting service The number of rows of medium-sized fruit trees, Indicates fig farmland awaiting service The length of the rectangle, This represents the average speed of the patrol drone during its spraying operation in the horizontal direction. Indicates patrol drone Fly from the station Arrive at the station The corresponding binary variable, It means arbitrary.

5. The three-dimensional path planning method for a fig seedling patrol drone according to claim 4, characterized in that, The cumulative work done by the patrol drone in uniform horizontal flight is expressed as follows: ； Indicates patrol drone Leaving the station The subsequent horizontal flight power.

6. The three-dimensional path planning method for a fig seedling patrol drone according to claim 4, characterized in that, The cumulative work done by the patrol drone during its operation is expressed as follows: ； Indicates the induced power factor. This indicates the rotational speed of the patrol drone's rotor. This indicates the rotor diameter of the patrol drone. This indicates the total mass of the patrol drone over time. A changing function, Indicates air density, This indicates the rotor disk area of ​​the patrol drone. Represents gravitational acceleration. This indicates the power of the patrol drone to overcome air resistance.

7. The three-dimensional path planning method for a fig seedling patrol drone according to claim 4, characterized in that, The cumulative work done by the patrol drone in its vertical take-off and landing state is expressed as follows: ； This represents the change in kinetic energy of the patrol drone. This represents the change in gravitational potential energy of the patrol drone. Indicates patrol drone When moving upwards or downwards, it does work by overcoming resistance.

8. The three-dimensional path planning method for a fig seedling patrol drone according to claim 1, characterized in that, The variable neighborhood search algorithm with utilization evaluation mechanism is used to solve the three-dimensional path planning model to obtain the final UAV service path set, including: S1031. Based on the planning parameters, generate an initial set of UAV service paths as the current solution, and record the current solution as the current optimal solution; S1032. Under the guidance of the utilization evaluation mechanism, the variable neighborhood search algorithm is used to perform variable neighborhood search processing on the current solution to generate candidate solutions; S1033. If the objective function value of the candidate solution is better than the objective function value of the current optimal solution, then update the current solution and the current optimal solution to the candidate solution; wherein, the objective function value of the candidate solution is the function value corresponding to the substitution of the candidate solution into the three-dimensional path planning model; the objective function value of the current optimal solution is the function value corresponding to the substitution of the current optimal solution into the three-dimensional path planning model; S1034. Repeat steps S1032 and S1033 until the number of consecutive iterations without improvement reaches the perturbation initiation threshold. S1035. When S1034 is true, perform a perturbation operation on the current optimal solution, and set the perturbed solution obtained after the perturbation operation as the current solution; S1036. Repeat steps S1032 to S1035 until a preset termination condition is met, and take the current optimal solution corresponding to the preset termination condition as the final UAV service path set.

9. The three-dimensional path planning method for a fig seedling patrol drone according to claim 8, characterized in that, The preset termination conditions are: the total running time of the solution process reaches a preset duration threshold, the number of iterations reaches an iteration number threshold, or the number of iterations in which the objective function value of the current optimal solution has not been improved for a continuous period of time reaches a preset stagnation threshold.

10. The three-dimensional path planning method for a fig seedling patrol drone according to claim 1, characterized in that, The utilization rate indicator is expressed as: ； in, This represents the value of the utilization rate indicator. Indicates the weighting coefficient. Indicates capacity utilization rate. Indicates power utilization rate. This indicates the ideal utilization rate of the medicine box capacity and the drone's battery power. ； ； Indicates fig farmland awaiting service , , Indicates fig farmland awaiting service , , This represents a collection of fig farms awaiting service. Indicates patrol drone From fig farms awaiting service Fly to fig farms awaiting service The corresponding binary variable, Indicates fig farmland awaiting service Required volume of medicine solution This indicates the maximum payload capacity of the patrol drone. This indicates the total number of patrol drones. Indicates patrol drone The estimated cumulative work done to complete the flight path. This indicates the maximum power of the patrol drone. Indicates patrol drone , , This indicates a collection of patrol drones. It means arbitrary.

Images