Multi-plant-protection unmanned aerial vehicle hybrid grey wolf optimization path planning method and control terminal

Abstract

This invention provides a hybrid gray wolf optimal path planning method and control terminal for multiple agricultural drones, relating to the field of agricultural drone control technology. The method includes: constructing an objective function for agricultural drone path planning by integrating path cost, altitude cost, turning angle cost, and collision avoidance cost; constructing a gray wolf algorithm model, setting a maximum number of iterations, initializing the population using SPM chaotic mapping, calculating the fitness values ​​of the gray wolf pack, and sorting the fitness values ​​in ascending order to obtain the top three wolves; updating the nonlinear convergence factor based on the current iteration number, updating the gray wolf positions based on the positions of the wolves, the nonlinear convergence factor, and the shrinking region function, and calculating the fitness value of the gray wolves after the position update; re-determining the top three wolves based on the updated fitness values ​​and outputting the optimal individual at the end of the iteration. This invention can improve pesticide spraying efficiency.

Description

Technical Field

[0001] This invention relates to the field of agricultural drone control technology, and in particular to a hybrid gray wolf optimization path planning method and control terminal for multiple agricultural drones. Background Technology

[0002] Agricultural drones can be used to monitor soil moisture and crop health in planting areas, as well as for intelligent irrigation and precision fertilization. They enable timely and effective crop protection and management, thereby improving agricultural production efficiency, reducing labor costs, minimizing environmental impact, and making a significant contribution to sustainable agricultural development. Path planning plays a crucial role in the operation of agricultural drones. Proper path planning can effectively improve flight efficiency, avoid redundant flights or missed areas, and thus save time and energy. For example, using agricultural drones to spray pesticides ensures even application, avoiding localized over- or under-application, and improving the quality of the operation. Furthermore, path planning helps agricultural drones avoid obstacles, ensuring safety during operations. The path planning problem for agricultural drones refers to designing effective operational routes to ensure that agricultural drones can effectively cover designated areas and perform tasks such as pesticide spraying, sowing, and crop monitoring. Path planning involves not only determining the start and end points of the flight but also considering various factors such as terrain, obstacles, the shape and size of the farmland, wind speed, flight speed, and energy consumption. The goal of path planning is to optimize the operational path of agricultural drones so that they can complete their tasks in the shortest time with the lowest energy consumption.

[0003] The Grey Wolf Optimization Algorithm (GLOO) effectively avoids local optima due to its powerful global search capability, making it suitable for solving complex path planning problems and exhibiting significant advantages in this area. However, in the path planning problem of targeted pesticide spraying by agricultural drones, the high dimensionality of the problem leads to a large search space, significantly reducing the convergence speed. Furthermore, the GLOO's search mechanism relies on group cooperation and hierarchical structure, making it prone to getting trapped in local optima during iteration. Especially in complex agricultural environments with numerous obstacles and constraints, the algorithm may fail to effectively escape local optima and find the globally optimal path. Additionally, during iteration, the population diversity of the GLOO may gradually decrease, particularly when the solution space is large, weakening the algorithm's global search capability. In agricultural drone path planning, this may hinder the algorithm from effectively exploring new paths, thus missing better ones. Therefore, a method is needed to improve the accuracy of targeted pesticide spraying by multiple agricultural drones while achieving effective coverage of the sprayed area. Summary of the Invention

[0004] This invention provides a hybrid gray wolf optimized path planning method and control terminal for multiple agricultural drones to solve the problems of precision and effective coverage of pesticide spraying at fixed points by multiple agricultural drones.

[0005] In a first aspect, embodiments of the present invention provide a hybrid gray wolf optimization path planning method for multiple agricultural drones, including: By considering path cost, altitude cost, turning angle cost, and collision avoidance cost, an objective function for path planning of agricultural drones is constructed. Flight trajectory planning is then performed for each agricultural drone based on this objective function. The flight trajectory planning steps include: A gray wolf algorithm model is constructed, a maximum number of iterations is set, the population is initialized using the SPM chaotic mapping, the fitness values ​​of the gray wolf pack are calculated, and the fitness values ​​are sorted in ascending order to obtain the top three. Wolf, wolves and Wolf; Update the nonlinear convergence factor based on the current iteration number, according to Wolf, wolves and The wolf's position, nonlinear convergence factor, and shrinking region function are used to update the gray wolf's position, and the humidity value of the gray wolf after the position update is calculated. Based on the updated Gray Wolves' fitness rating, the top three have been re-ranked. Wolf, wolves and Wolf; If the maximum number of iterations has been reached, the iteration ends, the optimal individual is output, and the final minimum fitness value of the UAV flight trajectory is obtained; otherwise, the operation of updating the nonlinear convergence factor based on the current number of iterations and subsequent operations is performed.

[0006] In one possible implementation, the objective function is: in, These are path cost, height cost, turning angle cost, and collision avoidance cost; The weighting coefficient for each individual agricultural drone.

[0007] In one possible implementation, The path cost, height cost, turning angle cost, and collision avoidance cost correspond to the following functional forms: in, These are functions corresponding to path cost, height cost, turning angle cost, and collision avoidance cost, respectively. It refers to the number of agricultural drones; It is the Euclidean distance between two path points; It is the first Agricultural drones were deployed in the first The coordinates of the path points; It is the number of path points; and Represents the maximum and minimum altitude during flight; It is the first The flight altitude of agricultural drones; It is the first Agricultural drones were deployed in the first The turning angle of a path point is expressed as: and All are finite numbers. , They represent the first frame, first Agricultural drones were deployed in the first The position at the next iteration; It is the first plant protection drones and the first Safe distance between agricultural drones.

[0008] In one possible implementation, updating the nonlinear convergence factor based on the current iteration number includes: The nonlinear convergence factor is updated using the following formula: in, It is the maximum value of the convergence factor; The number of iterations for the Grey Wolf optimization algorithm; This represents the maximum number of iterations for the Grey Wolf optimization algorithm. is a positive integer.

[0009] In one possible implementation, the shrinking region function is: in, Indicates the first In the nth iteration The optimal position of the dimension. Indicates the first A random position in dimension, To reduce the shrinkage perturbation of the region function.

[0010] In one possible implementation, the basis Wolf, wolves and The wolf's position is updated using the wolf's location, nonlinear convergence factor, and shrinking region function, including: Update the position and update coefficients based on the nonlinear convergence factor. and ; according to Wolf, wolves and The wolf's position is calculated using the following formula, yielding three sets of optimal candidate positions: , , , in, , , The optimal position for the candidate; , and For the first During the next iteration Wolf, wolves and The wolf's location; This indicates the current position of the Grey Wolves. Combining the shrinking region function, the position of the gray wolf is updated using the position update formula; the position update formula is: in, It is obedience Random numbers with a mean square distribution; , ( ) is the first During the nth iteration, the 1st The first time to set up an agricultural drone Dimensional position; This is a function for narrowing the region.

[0011] In one possible implementation, the position update coefficients and The calculation formulas are as follows: in, and For interval Random variables on.

[0012] In one possible implementation, the initialization of the population using the SPM chaotic mapping includes: The initial population is generated using the SPM chaotic mapping formula, which is: in, It is a random number; For the first The first agricultural drone The output value of the submapping.

[0013] In one possible implementation, calculating the fitness value of the gray wolf pack includes: The fitness value of the gray wolf pack is calculated by substituting the position of the gray wolf into the objective function.

[0014] In a second aspect, embodiments of the present invention provide a control terminal, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect or any possible implementation of the first aspect.

[0015] In this embodiment of the invention, an objective function is constructed by comprehensively considering path cost, altitude cost, turning angle cost, and collision avoidance cost. This comprehensively considers key influencing factors in the path planning of multiple agricultural drones, ensuring that the planned path balances efficiency and safety. The population is initialized using SPM chaotic mapping to maintain population diversity; this increases the exploratory nature of the gray wolf optimization algorithm, improves search efficiency, reduces the risk of early convergence to local optima, increases the chance of escaping local optima, and accelerates the algorithm's convergence speed. By updating the nonlinear convergence factor and combining it with a shrinking region function to update the gray wolf position, the algorithm's exploration and development capabilities can be dynamically adjusted. The path is continuously optimized during iteration, and the search range is gradually reduced in the later stages of the search, focusing on a more refined search within the current optimal region. This leads to the construction of a new position update formula for the gray wolf optimization algorithm during the hunting phase. By iteratively determining the maximum number of iterations to output the optimal trajectory, efficient planning of multiple agricultural drone paths is achieved. This effectively solves the problems of slow convergence and susceptibility to local optima in traditional path planning. It is suitable for large-scale agricultural operations in complex farmland environments, provides reasonable path schemes for collaborative operations of multiple agricultural drones, improves the accuracy of targeted pesticide spraying by multiple agricultural drones, and achieves effective coverage of the sprayed area. Attached Figure Description

[0016] Figure 1 This is an application scenario diagram of the hybrid gray wolf optimized path planning method for multiple agricultural drones provided in this embodiment of the invention; Figure 2 This is a flowchart illustrating the implementation of the hybrid gray wolf optimized path planning method for multiple agricultural drones provided in this embodiment of the invention. Figure 3 This is a comparison chart of the nonlinear convergence factor and the linear convergence factor in this invention; Figure 4This is a convergence comparison diagram of the hybrid gray wolf optimization algorithm in this invention with four other optimization algorithms; Figure 5 This is a side view comparison diagram of the paths planned by the hybrid gray wolf optimization algorithm and four other optimization algorithms in this invention; Figure 6 This is a top-down comparison diagram of the paths planned by the hybrid gray wolf optimization algorithm and four other optimization algorithms in this invention; Figure 7 This is a side view of the multi-drone path planning in the hybrid gray wolf optimization algorithm of this invention; Figure 8 This is a top view of the multi-agricultural drone path planned by the hybrid gray wolf optimization algorithm in this invention; Figure 9 This is a map showing the spraying path of the agricultural drone planned by the hybrid gray wolf optimization algorithm in this invention. Detailed Implementation

[0017] This invention proposes a hybrid gray wolf optimization path planning method for multiple agricultural drones. This hybrid optimization algorithm is based on the increase of population diversity, the design of nonlinear convergence factors, and the construction of shrinkage region functions, which solves the problems of fixed-point pesticide spraying and mulching spraying of multiple agricultural drones.

[0018] The embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0019] Figure 1 This diagram illustrates an application scenario of the hybrid gray wolf optimal path planning method for multiple agricultural drones provided in this embodiment of the invention. Figure 1 As shown, multiple agricultural drones work simultaneously to spray pesticides on crops. These drones are connected to a path planning terminal.

[0020] The path planning terminal centrally acquires crop planting information and information from multiple agricultural drones. It integrates information such as terrain, obstacles, shape and size of farmland, wind speed, flight speed and energy consumption to model a three-dimensional planning space, sets the number, shape, location, size and height of obstacles in the working environment, and sets the operational parameters of agricultural drones to complete the flight path planning for multiple agricultural drones.

[0021] In practice, operators can use a path planning terminal to view the paths of multiple agricultural drones and monitor real-time spraying operations. This path planning terminal can be a portable mobile device such as a laptop or mobile phone, or it can be a server, with operators using a terminal device connected to the server to view the paths of multiple agricultural drones and monitor real-time spraying operations.

[0022] Figure 2This is a flowchart illustrating the implementation of the hybrid gray wolf optimized path planning method for multiple agricultural drones provided in this embodiment of the invention. Figure 2 As shown, the method includes the following steps: S201, taking into account path cost, altitude cost, turning angle cost, and collision avoidance cost, constructs an objective function for path planning of agricultural drones, so as to plan the flight trajectory of each agricultural drone according to the objective function.

[0023] The execution subject in various embodiments of the present invention can be a server, processor, microprocessor, or other device with data processing capabilities. In actual implementation, the specific implementation method of the execution subject can be selected according to actual needs. This embodiment does not impose any particular limitation on this; any device with data processing capabilities is acceptable. The present invention... Figure 1 The path planning terminal shown is the main focus of the explanation.

[0024] Among these factors, path cost is directly related to the total flight distance of the drone. Reducing path cost can lower energy consumption and shorten operation time, meeting the core efficiency requirements of agricultural production. Altitude cost is used to constrain the drone's flight altitude within a reasonable range, avoiding collisions with obstacles due to excessively low altitude or affecting spraying accuracy due to excessively high altitude, thus ensuring operation quality. Turning angle cost is designed based on the drone's maneuverability. Excessive turning angles increase energy consumption for attitude adjustment and may lead to uneven spraying; cost constraints can make the path smoother. Collision avoidance cost is a key constraint for multi-drone collaborative operations. When multiple drones are operating simultaneously, a safe distance must be maintained to avoid collisions that could damage equipment or interrupt operations. By integrating these four types of costs, the objective function can comprehensively reflect the overall performance of the path, improve the efficiency of path planning for multiple agricultural drones, enhance the accuracy of targeted pesticide spraying by multiple agricultural drones, and achieve effective coverage of the sprayed area.

[0025] S202: Construct the gray wolf algorithm model, set the maximum number of iterations, use SPM chaotic mapping to initialize the population, calculate the fitness value of the gray wolf pack, sort the fitness values ​​in ascending order, and obtain the top three... Wolf, wolves and Wolf.

[0026] The Grey Wolf Optimization Algorithm was chosen to leverage its strong global search capability, making it suitable for complex optimization problems such as path planning. SPM chaotic mapping is used to initialize the population, addressing the insufficient diversity of traditional random initialization. The ergodicity of the chaotic mapping generates more uniformly distributed initial solutions, covering a wider potential path space and reducing the risk of the algorithm getting trapped in local optima.

[0027] Calculate fitness values ​​and filter Wolf, Wolf, The algorithm simulates the social hierarchy of wolf packs, using the three best-performing individuals in the group to guide the overall search direction and improve the convergence efficiency of the algorithm.

[0028] S203, update the nonlinear convergence factor according to the current iteration number, according to Wolf, wolves and The wolf's position, nonlinear convergence factor, and shrinking region function are used to update the gray wolf's position, and the humidity value of the gray wolf after the position update is calculated.

[0029] In one possible implementation, the fitness value of the gray wolf pack is calculated by: Substitute the positions of the gray wolves into the objective function to calculate the fitness value of the gray wolf pack.

[0030] The ability to explore and develop balancing algorithms for dynamically updating nonlinear convergence factors. For example... Figure 3 As shown, the blue line represents the nonlinear convergence factor provided in the embodiments of this application. In the early stages of iteration, the factor value is large and changes slowly, enhancing the global search scope to discover potential optimal regions; in the later stages of iteration, the factor value decreases, focusing on local regions for fine-tuning and avoiding ineffective searches.

[0031] By combining the position update function of the narrowing region, the search range is gradually narrowed during the iteration process, which improves the search accuracy of the algorithm when it is close to the optimal solution and reduces the waste of computing resources.

[0032] S204, Based on the updated Gray Wolf's fitness value, the top three rankings have been redefined. Wolf, wolves and Wolf.

[0033] Re-filter after each location update Wolf, Wolf, Wolves ensure that the population is always guided by the best individual and continues to evolve towards a better path.

[0034] S205: If the maximum number of iterations has been reached, the iteration ends, the optimal individual is output, and the drone flight trajectory with the final minimum fitness value is obtained; otherwise, the operation of updating the nonlinear convergence factor based on the current number of iterations and subsequent operations is performed.

[0035] Setting a maximum number of iterations as the termination condition avoids resource consumption caused by infinite algorithm iteration and ensures the timeliness of path planning through a fixed iteration period. When the iteration terminates, the flight trajectory corresponding to the optimal individual is output. This trajectory is the solution with the minimum objective function value, that is, the solution with the lowest comprehensive path cost, altitude cost, turning angle cost, and collision avoidance cost. It can be directly used to guide the collaborative operation of multiple agricultural drones.

[0036] In actual implementation, after obtaining the final minimum fitness value of the UAV flight trajectory, the process also includes generating a 3D view and a top-down path map for actual flight execution.

[0037] In this embodiment, an objective function is constructed by comprehensively considering path cost, altitude cost, turning angle cost, and collision avoidance cost. This allows for a comprehensive consideration of key influencing factors in multi-drone path planning, ensuring that the planned path balances efficiency and safety. The population is initialized using SPM chaotic mapping to maintain population diversity; this increases the exploratory nature of the gray wolf optimization algorithm, improves search efficiency, reduces the risk of early convergence to local optima, increases the chance of escaping local optima, and accelerates the algorithm's convergence speed. By updating the nonlinear convergence factor and combining it with a shrinking region function to update the gray wolf position, the algorithm's exploration and development capabilities can be dynamically adjusted. The path is continuously optimized during iteration, and the search range is gradually reduced in the later stages of the search, focusing on a more refined search within the current optimal region. This leads to a new position update formula for the gray wolf optimization algorithm during the hunting phase. By iteratively determining the maximum number of iterations to output the optimal trajectory, efficient planning of multiple agricultural drone paths is achieved. This effectively solves the problems of slow convergence and susceptibility to local optima in traditional path planning. It is suitable for large-scale agricultural operations in complex farmland environments, provides reasonable path schemes for collaborative operations of multiple agricultural drones, improves the accuracy of targeted pesticide spraying by multiple agricultural drones, and achieves effective coverage of the sprayed area.

[0038] The above provides an overview of the overall scheme. The following embodiments provide specific limitations on the objective function, nonlinear convergence factor, shrinkage region function, and the generation of the initial population through the SPM chaotic mapping formula.

[0039] In one possible implementation, the objective function is: in, These are path cost, height cost, turning angle cost, and collision avoidance cost; The weighting coefficient for each individual agricultural drone.

[0040] In different embodiments, the weight of each cost is determined based on specific needs such as prioritizing energy saving or prioritizing safety, thereby adjusting the cost coefficients in the above objective function.

[0041] In this embodiment, by introducing a cost coefficient to weight and integrate path cost, altitude cost, turning angle cost, and collision avoidance cost, the importance of each cost factor in path planning can be flexibly adjusted according to actual operational needs. This makes the objective function more closely match the actual operational scenarios of multiple agricultural drones, thereby making the path planned based on the objective function more targeted. Under different operational needs, an optimized path that meets expectations can be obtained, improving the flexibility and practicality of path planning.

[0042] In one possible implementation, the corresponding functional forms of path cost, height cost, turning angle cost, and collision avoidance cost are as follows: in, These are functions corresponding to path cost, height cost, turning angle cost, and collision avoidance cost, respectively. It refers to the number of agricultural drones; It is the Euclidean distance between two path points; It is the first Agricultural drones were deployed in the first The coordinates of the path points; It is the number of path points; and Represents the maximum and minimum altitude during flight; It is the first The flight altitude of agricultural drones; It is the first Agricultural drones were deployed in the first The turning angle of a path point is expressed as: and All are finite numbers. , They represent the first frame, first Agricultural drones were deployed in the first The position at the next iteration; It is the first plant protection drones and the first Safe distance between agricultural drones.

[0043] In practical implementation, to ensure that agricultural drones can efficiently complete spraying tasks within a limited time, their flight path should be as short as possible. The entire flight path is divided into... The total flight distance is calculated using a number of path segments. The flight distance cost of the agricultural drone is defined as: in, This refers to the number of drones. For the first The drone in The coordinates of the path points.

[0044] When agricultural drones fly too high or too low, energy consumption increases. Furthermore, flying too low may increase the risk of collision between the drone and crops. Therefore, it is essential to ensure that the relative altitude between the drone and crops is greater than or equal to a specific value to guarantee flight safety. This invention imposes the following constraints on the drone's flight altitude: Considering the limitations of drone maneuverability, the planned path should remain smooth. Larger turning angles may require more energy for the drone to adjust its flight attitude and could cause the agricultural drone to deviate from the spraying area, thus increasing energy consumption and reducing flight time and operational efficiency. Therefore, this invention imposes the following constraints on turning angles: , in, For the first The drone in Turning angles at each path point It is a finite number. The expression is: Simultaneous operation of multiple agricultural drones can effectively improve operational efficiency. During operation, it is crucial to ensure that drones do not collide. Therefore, maintaining a safe distance between drones is essential. To avoid collisions, this invention defines collision constraints between drones as follows: Based on the above considerations, the following path planning model for multiple agricultural drones is established: in, The weights represent the cost function.

[0045] This involves creating a 3D planning space model based on the actual environment, setting the number, shape, location, size, and height of obstacles in the operational environment. It also includes setting operational parameters for the agricultural drones, including the number of drones, the start and end coordinates for each drone, the number of waypoints, and the maximum flight altitude. Minimum flight altitude Maximum turning angle Minimum safe distance between agricultural drones and other agricultural drones Set the maximum number of iterations for the algorithm. Population size The maximum value of the convergence factor and shrinkage perturbation of shrinkage region function .

[0046] In this embodiment, the path cost function, by calculating the sum of Euclidean distances between path points, accurately quantifies the travel costs of the UAV's flight, providing a clear optimization direction for shortening flight distances. The altitude cost function, combined with maximum and minimum flight altitude constraints, avoids increased energy consumption or collision risks caused by improper UAV flight altitude. The turning angle cost function, by limiting the maximum turning angle, ensures the smoothness of the UAV's flight path, reducing energy consumption and operational deviations during attitude adjustments. The collision avoidance cost function, constructed based on the safe distance between UAVs, effectively avoids collisions during multi-UAV collaborative operations. The cooperation of these cost functions makes the calculation of the objective function more accurate, providing a reliable cost assessment basis for subsequent algorithm iteration and optimization, and further improving the safety of path planning.

[0047] In one possible implementation, population initialization is performed using the SPM chaotic mapping, including: The initial population is generated using the SPM chaotic mapping formula, which is: in, It is a random number; For the first The first agricultural drone The output value of the submapping.

[0048] The sequences generated by the SPM chaotic mapping exhibit better randomness, producing an initial population with a uniform distribution and good diversity, thus avoiding the population concentration phenomenon that may occur in traditional random initialization. Therefore, this invention employs the SPM chaotic mapping for population initialization. Initializing the population through the SPM chaotic mapping can generate a uniformly distributed and comprehensive initial solution set, at which point the number of iterations is reduced. .

[0049] In this embodiment, the chaotic mapping formula combines random numbers with piecewise functions to generate an initial population that is evenly distributed and has good diversity. Compared with the traditional random initialization method, it can avoid the problem of limited search range caused by excessive concentration of the initial population, so that the initial population can more comprehensively cover the search space, provide richer initial solutions for subsequent iterations of the algorithm, help the algorithm to quickly find potential better regions in the early stage of iteration, reduce the risk of the algorithm getting stuck in local optima, and improve the iteration speed of multi-agricultural drone path planning.

[0050] In one possible implementation, the nonlinear convergence factor is updated based on the current iteration number, including: The nonlinear convergence factor is updated using the following formula: in, It is the maximum value of the convergence factor; The number of iterations for the Grey Wolf optimization algorithm; This represents the maximum number of iterations for the Grey Wolf optimization algorithm. is a positive integer.

[0051] like Figure 3 As shown, the red dashed line represents the linear convergence factor in the traditional Grey Wolf optimization algorithm. The blue solid line represents the nonlinear convergence factor provided in this embodiment. In the traditional Grey Wolf optimization algorithm, the convergence factor decreases linearly from 2 to 0. This linear decrease limits the adaptability and search capability of the individual, making it difficult to flexibly cope with the constantly changing requirements during global and local searches, which may lead to the algorithm getting trapped in local optima too early or having a slow convergence speed. In this embodiment, the use of a nonlinear convergence factor can enhance the adaptability of the algorithm, accelerate the convergence speed, reduce computation time, and help avoid getting trapped in local optima in high-dimensional problems, thereby achieving a more efficient search. Therefore, this invention designs the above-mentioned nonlinear convergence factor.

[0052] In this embodiment, compared to the traditional linear convergence factor, this nonlinear formula adjusts the convergence factor through the exponential relationship between the number of iterations and the maximum number of iterations. In the early stages of iteration, it maintains a larger convergence factor value, enhancing the algorithm's global exploration capability and helping it quickly traverse the search space to find potential optimal solutions. In the later stages of iteration, the convergence factor value gradually decreases, allowing the algorithm to focus more on local development and finely optimize the found better regions. This avoids an imbalance between exploration and development during iteration, effectively improving the algorithm's convergence speed and accuracy, and ensuring that multi-drone path planning can efficiently find the optimal solution.

[0053] In one possible implementation, the position update coefficients and The calculation formulas are as follows: in, and For interval Random variables on.

[0054] In actual implementation, the updated convergence factor is used. calculate and To update coefficients based on the updated position and Optimize the position of the gray wolf.

[0055] In this embodiment, by introducing a random variable in the interval [0,1], the coefficients are made random. This randomness increases the diversity of the gray wolf population's location updates, avoids the over-concentration of individual locations during the iteration process, thereby maintaining population diversity, enhancing the algorithm's global exploration capability, and reducing the probability of the algorithm getting stuck in local optima. At the same time, the coefficients are associated with the nonlinear convergence factor, which enables the coefficients to be dynamically adjusted with the iteration process, adapting to the exploration and development needs of the algorithm at different iteration stages, further improving the rationality of location updates, and optimizing the paths of multiple agricultural drones.

[0056] In one possible implementation, according to Wolf, wolves and The wolf's position is updated using the wolf's location, nonlinear convergence factor, and shrinking region function, including: Update the position and update coefficients based on the nonlinear convergence factor. and ; according to Wolf, wolves and The wolf's position is calculated using the following formula, yielding three sets of optimal candidate positions: , , , in, , , The optimal position for the candidate; , and For the first During the next iteration Wolf, wolves and The wolf's location; This indicates the current position of the Grey Wolves. Combining the shrinking region function, the position of the gray wolf is updated using the position update formula; the position update formula is: in, It is obedience Random numbers with a mean square distribution; , ( ) is the first During the nth iteration, the 1st The first time to set up an agricultural drone Dimensional position; This is a function for narrowing the region.

[0057] Based on the aforementioned embodiments, the current optimal solution for the gray wolf's position can be expressed as: In this embodiment, the position update coefficients are first updated based on the nonlinear convergence factor, and then... Wolf, wolves and The wolf position calculation uses three sets of candidate optimal positions, which can make full use of the guiding role of the best individual in the population to provide a reliable direction for position updates. Then, the gray wolf position is updated by combining the shrinking region function and the position update formula. This process not only retains the guiding advantage of the best individual, but also enhances the fineness of the local search through the shrinking region function, so that the gray wolf population can continuously move towards a better position in the iteration, avoid the population getting stuck in local optima, speed up the convergence speed of the algorithm, and ensure that the paths of multiple agricultural drones can be quickly iterated to the optimal state.

[0058] In one possible implementation, the shrinking region function is: in, Indicates the first In the nth iteration The optimal position of the dimension. Indicates the first A random position in dimension, To reduce the shrinkage perturbation of the region function.

[0059] In practical implementation, in order to overcome the problem of slow convergence speed in the later stages of the traditional gray wolf optimization algorithm, a shrinking region function was designed. This function can gradually reduce the search range in the later stages of the search and concentrate on a more refined search in the current optimal region.

[0060] In this embodiment, the shrinking region function, combined with the optimal and random positions during the iteration process, controls the search range by shrinking the perturbation. In the later stages of the algorithm iteration, the search area can be gradually reduced, allowing the algorithm to focus its efforts on the current optimal area for a more refined search. This avoids wasting computational power in large-scale searches and reduces the problem of missing optimal solutions due to an excessively large search range. It further enhances the algorithm's local search capability in the later stages of iteration, helping the algorithm to find the global optimal solution for multiple agricultural drone paths more accurately, improving the accuracy of path planning, and effectively achieving fixed-point pesticide spraying control and effective coverage of the overall spraying area by multiple drones.

[0061] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0062] To better demonstrate the advantages of the proposed Mix Grey Wolf Optimizer (MGWO), the Grey Wolf Optimizer (GWO), Whale Optimization Algorithm (WOA), Catch Fish Optimization Algorithm (CFOA), and African Vultures Optimization Algorithm (AVOA) are compared with the proposed Mix Grey Wolf Optimizer (MGWO). Each algorithm was run 50 times under the same parameter conditions, and the results are shown in Table 1. Figure 4 As shown. Compared with other optimization algorithms, MGWO obtained the minimum objective function value, the minimum mean value, and a good standard deviation among the compared algorithms, indicating that the optimization algorithm proposed in this invention has better optimization capabilities. From Figure 4 As can be seen, in the early stages of iteration, MGWO's descent rate is faster than other optimization algorithms, indicating that it can find a better solution more quickly. At the end of the iteration, MGWO reaches the minimum optimal fitness value, indicating that it found the optimal or near-optimal solution in this test. Furthermore, MGWO's curve is relatively smooth, without significant fluctuations, indicating that the optimization algorithm exhibits good stability during the iteration process.

[0063] Furthermore, the proposed hybrid gray wolf optimization algorithm is compared with four other optimization algorithms in three-dimensional UAV path planning. The population size and maximum number of iterations are both set to 300. The flight start point is [20, 20, 20] meters, and the end point is [60, 140, 200] meters. The path planning comparison results are as follows: Figure 5 and Figure 6 As shown in the figure, in this 3D UAV path planning, all five algorithms were able to avoid obstacles and successfully reach the destination, but the path generated by MGWO was smoother and shorter. Therefore, the hybrid gray wolf algorithm proposed in this invention can be effectively applied to 3D UAV path planning.

[0064] Table 1 Algorithm Comparison Results To further verify the effectiveness of the proposed hybrid gray wolf optimization algorithm in path planning for multiple agricultural drones, it was applied to the fixed-point spraying path planning problem of multiple agricultural drones in a mountainous environment. The spatial data of the terrain environment where three identical agricultural drones were located was 1000m × 1000m × 800m, with nine obstacles set in the environment. The starting coordinates were [150, 200, 100]m, [400, 100, 100]m, and [200, 100, 100]m, and the ending coordinates were [800, 800, 200]m, [900, 600, 150]m, and [800, 700, 150]m, respectively. The maximum flight altitude of the agricultural drones was [not specified]. =200 meters, minimum flight altitude 50 meters, maximum turning angle Minimum safe distance between agricultural drones 5 meters. Set the maximum number of iterations for the algorithm. 300 times, population size 500, the maximum value of the convergence factor. 2. Shrinking perturbation of the region function The weights of the cost function are respectively .

[0065] Figure 7 and Figure 8 The specific paths of three agricultural drones are shown. Visual analysis of the path maps leads to the conclusion that, due to the large number of obstacles and complex constraints, a detour strategy should be adopted as much as possible during the path planning process to avoid collisions caused by obstacle constraints. The hybrid gray wolf optimization algorithm designed in this invention can best fit the actual terrain and find the optimal path planning scheme in a timely manner within the safe area of ​​complex obstacles, providing an important foundation for agricultural scheduling in real-world scenarios.

[0066] To verify the feasibility of the proposed hybrid gray wolf optimization algorithm in planning agricultural drone spraying operations, this embodiment selects an irregular quadrilateral plot for testing. The total area of ​​the farmland is 3800 square meters, and the plot coordinates, arranged counterclockwise, are (10, 5) meters, (20, 70) meters, (70, 80) meters, and (80, 20) meters. Obstacle information is shown in Table 2. The drone flies at a height 2 meters above the crop leaf tips, with a spraying width of 4 meters and a safe distance of 5 meters from obstacles. The coverage path planned using the proposed hybrid gray wolf optimization algorithm is shown below. Figure 9As shown, the total flight distance of the agricultural drone was 1155.4 meters, the effective farmland area was 2928.25 square meters, and the actual coverage area was 2896.5 square meters, with a coverage rate as high as 97.91%. It can be seen that the hybrid gray wolf optimized path planning algorithm proposed in this invention has successfully completed the spraying task while safely avoiding all obstacles, which meets the actual needs of agricultural drones in spraying operations.

[0067] Table 2 Obstacle Location Information The following are device embodiments of the present invention. For details not described in detail, please refer to the corresponding method embodiments described above.

[0068] This invention also provides a control terminal, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the methods described in the above method embodiments. Exemplarily, the control terminal can be a smartphone, tablet computer, laptop computer, desktop computer, smart speaker, smartwatch, etc., and is not limited thereto.

[0069] In the above embodiments, the descriptions of each embodiment have their own emphasis. Parts not detailed or described in a particular embodiment can be referred to in the relevant descriptions of other embodiments. Unless otherwise specified or in conflict with logic, the terminology and / or descriptions between different embodiments are consistent and can be referenced interchangeably. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships.

[0070] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A hybrid gray wolf optimization path planning method for multiple agricultural drones, characterized in that, include: By considering path cost, altitude cost, turning angle cost, and collision avoidance cost, an objective function for path planning of agricultural drones is constructed, and flight trajectory planning is performed for each agricultural drone based on the objective function. Flight trajectory planning steps include: A gray wolf algorithm model is constructed, a maximum number of iterations is set, the population is initialized using the SPM chaotic mapping, the fitness values ​​of the gray wolf pack are calculated, and the fitness values ​​are sorted in ascending order to obtain the top three. Wolf, wolves and Wolf; Update the nonlinear convergence factor based on the current iteration number, according to Wolf, wolves and The wolf's position, nonlinear convergence factor, and shrinking region function are used to update the gray wolf's position, and the humidity value of the gray wolf after the position update is calculated. Based on the updated Gray Wolves' fitness rating, the top three have been re-ranked. Wolf, wolves and Wolf; If the maximum number of iterations has been reached, the iteration ends, the optimal individual is output, and the final minimum fitness value of the UAV flight trajectory is obtained; otherwise, the operation of updating the nonlinear convergence factor based on the current number of iterations and subsequent operations is performed. The step of updating the nonlinear convergence factor based on the current iteration number includes: The nonlinear convergence factor is updated using the following formula: in, It is the maximum value of the convergence factor; The number of iterations for the Grey Wolf optimization algorithm; This represents the maximum number of iterations for the Grey Wolf optimization algorithm. is a positive integer; The region reduction function is: in, Indicates the first In the nth iteration The optimal position of the dimension. Indicates the first A random position in dimension, To reduce the shrinkage perturbation of the region function; According to Wolf, wolves and The wolf's position is updated using the wolf's location, nonlinear convergence factor, and shrinking region function, including: Update the position and update coefficients based on the nonlinear convergence factor. and ; according to Wolf, wolves and The wolf's position is calculated using the following formula, yielding three sets of optimal candidate positions: ， ， ， in, , , The optimal position for the candidate; , and For the first During the next iteration Wolf, wolves and The wolf's location; This indicates the current position of the Grey Wolves. Combining the shrinking region function, the position of the gray wolf is updated using the position update formula; the position update formula is: in, It is obedience Random numbers with a mean square distribution; , It is the first During the nth iteration, the 1st The first time to set up an agricultural drone Dimensional position; This is a function for narrowing the region.

2. The hybrid gray wolf optimal path planning method for multiple agricultural drones according to claim 1, characterized in that, The objective function is: in, These are path cost, height cost, turning angle cost, and collision avoidance cost; , The weighting coefficient for each individual agricultural drone.

3. The hybrid gray wolf optimal path planning method for multiple agricultural drones according to claim 1 or 2, characterized in that, The path cost, height cost, turning angle cost, and collision avoidance cost correspond to the following functional forms: in, These are functions corresponding to path cost, height cost, turning angle cost, and collision avoidance cost, respectively. It refers to the number of agricultural drones; It is the Euclidean distance between the two path points; It is the first Agricultural drones were deployed in the first The coordinates of the path points; It is the number of path points; and Represents the maximum and minimum altitude during flight; It is the first The flight altitude of agricultural drones; It is the first Agricultural drones were deployed in the first The turning angle of a path point is expressed as: and All are finite numbers; , They represent the first frame, first Agricultural drones were deployed in the first The position at the next iteration; It is the first plant protection drones and the first Safe distance between agricultural drones.

4. The hybrid gray wolf optimal path planning method for multiple agricultural drones according to claim 1, characterized in that, Location update coefficient and The calculation formulas are as follows: in, and For interval Random variables on.

5. The hybrid gray wolf optimal path planning method for multiple agricultural drones according to claim 1, characterized in that, The initialization of the population using SPM chaotic mapping includes: The initial population is generated using the SPM chaotic mapping formula, which is: in, It is a random number; For the first The first agricultural drone The output value of this mapping.

6. The hybrid gray wolf optimal path planning method for multiple agricultural drones according to claim 1, characterized in that, The calculation of the fitness value of the gray wolf pack includes: The fitness value of the gray wolf pack is calculated by substituting the position of the gray wolf into the objective function.

7. A control terminal, characterized in that, It includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method as described in any one of claims 1 to 6.

Images