A method for flight path planning of agricultural UAVs based on an improved Harris Eagle algorithm

Abstract

This invention relates to the field of path planning technology, and particularly to a method for agricultural UAV trajectory planning based on an improved Harris Eagle algorithm. The method includes: constructing a site environment model; determining an objective function based on the site environment model; initializing parameters; initializing the population using a Chebyshev mapping, initializing individuals, and retaining Harris Eagle individuals with higher fitness to form the final initial population; executing a global exploration strategy or a local development strategy; continuously updating the current global optimal solution by comparing the fitness values ​​of all individuals in the entire search space, and outputting the optimal individual fitness value and position when the stopping condition is met. This invention combines the Chebyshev mapping, a nonlinear escape energy factor mechanism, and a back-learning cross-operator perturbation strategy. By constructing a trajectory planning model that integrates constraints such as the UAV flight area and maximum climb angle, a safe, reliable, and low-cost UAV trajectory planning scheme is obtained.

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, and in particular relates to a method for planning the flight path of agricultural unmanned aerial vehicles based on an improved Harris Eagle algorithm. Background Technology

[0002] Subtropical fruits are a pillar industry of agriculture in western Guangdong, with the region primarily consisting of mountainous and hilly areas. Severe outbreaks of pests and diseases can have a catastrophic impact on fruit safety, directly leading to a significant decline in fruit quality and yield. In recent years, with the rapid development of emerging technologies such as the Internet of Things and artificial intelligence, agricultural drones have been widely used in agricultural pesticide application. Compared to manual pesticide application, using drones for crop pest and disease control offers technological advantages. Therefore, it is necessary to plan an efficient flight path in the known complex terrain of mountain orchards, taking into account constraints such as the flight area and maximum climb angle of the drones used for pesticide application, in order to improve flight safety and reduce flight costs.

[0003] Many scholars have used artificial intelligence algorithms for path planning of unmanned aerial vehicles (UAVs). Among them, metaheuristic algorithms, as a novel type of heuristic algorithm, have been widely used due to their advantages such as not relying on specific conditions of a particular problem, strong global optimization capabilities, and few parameter settings. Existing technologies propose a hybrid metaheuristic algorithm for complex unknown 3D environments, combining the finite difference method with the cuckoo algorithm, which accelerates the global convergence speed while maintaining strong robustness. Existing technologies also propose an artificial bee colony optimization algorithm based on chaotic operators, establish a UAV flight model, and verify the effectiveness, feasibility, and robustness of the optimization algorithm.

[0004] There is considerable research on UAV path planning algorithms, but most results focus on improving the algorithm model, and two problems exist: First, even after performance improvements, the strategies for enhancing the algorithm to escape local optima are not ideal. In particular, in methods using swarm optimization, the chaotic characteristics of different mapping strategies differ during the population initialization phase, and the balance of linearly changing factors is weak in balancing the algorithm's exploration and development capabilities. Second, the flight path of UAVs for applying pesticides to orchards in mountainous areas needs to be planned in conjunction with constraints such as the terrain environment of hilly orchards and the maximum climb angle. Summary of the Invention

[0005] The purpose of this invention is to provide a flight path planning method for agricultural unmanned aerial vehicles based on an improved Harris Eagle algorithm, aiming to solve the problem that the existing technology has unsatisfactory strategies for enhancing the algorithm to escape local optima after the algorithm performance is improved.

[0006] This invention is implemented as follows: a method for planning the flight path of an agricultural unmanned aerial vehicle (UAV) based on an improved Harris Eagle algorithm, the method comprising:

[0007] Obtain the environmental conditions of the site, construct a site environment model, and determine the objective function based on the site environment model;

[0008] The objective function is solved by improving the Harris Eagle algorithm:

[0009] Initialization parameters, including population size Maximum number of iterations Dimensions of the objective function Initial value upper and lower bounds and ;

[0010] The population was initialized using Chebyshev mapping, and individuals were initialized. Individuals with higher fitness were retained to form the final initial population.

[0011] renew Value, calculate the prey's nonlinear escape energy factor ,according to The numerical values ​​can be used to execute a global exploration strategy or a local development strategy.

[0012] By comparing the fitness values ​​of all individuals in the entire search space, the current global optimal solution is continuously updated, and it is determined whether the maximum number of iterations T has been reached. If the condition is met, the fitness value and position of the optimal individual are output; otherwise, the population is reinitialized.

[0013] Preferably, in the steps of acquiring the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, the mathematical model of the site environment model is expressed as follows:

[0014] ;

[0015] In the formula, (x i y i Let h be the center coordinates of the i-th peak; i For terrain parameters, control the height, x si and y si These represent the attenuation amounts of the i-th peak along the x-axis and y-axis, respectively, controlling the slope, and n is the total number of peaks.

[0016] Preferably, in the steps of acquiring the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, terrain constraints are set, as follows:

[0017] ;

[0018] In the formula, Z i Z represents the drone's flight altitude, Z is the terrain function, and Z(x) i,y i ) represents (x i ,y i The altitude of the terrain will be used to determine the fitness level. When the flight altitude is less than or equal to the terrain altitude, the fitness level will be amplified by a preset ratio.

[0019] Preferably, in the steps of acquiring the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, location constraints are set, as follows:

[0020] ;

[0021] In the formula, x max y max and z max , where are the maximum values ​​in each dimension of the solution space, and n is a natural number. When a path point exceeds the solution space, the fitness will be amplified.

[0022] Preferably, in the steps of acquiring the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, the maximum climb angle constraint γ is set, expressed as:

[0023] ;

[0024] in, This is the maximum climb angle.

[0025] Preferably, the expression for the objective function is:

[0026] ;

[0027] In the formula: ;

[0028] ;

[0029] ;

[0030] Where c1, c2, and c3 are the control variables for terrain constraints, position constraints, and angle constraints θ, respectively. max Let P be the maximum constraint angle, and let P be the solution space of the optimization problem.

[0031] Preferably, in the step of initializing the population using Chebyshev mapping, initializing individuals, and retaining Harris eagles with higher fitness to form the final initial population, the initial position of the Harris eagle population is set using Chebyshev mapping. Therefore, the initial population position is:

[0032] ;

[0033] in, An individual randomly selected from the current population; and For the current Harris Hawks The second iteration and the first The position vector of the next iteration; This represents the current optimal position for the individual, i.e., the prey. The average position of individuals in the population; This represents the average positional difference between prey and individuals. and These are the lower and upper bounds of the population, respectively. For population size; Z1, Z 2、 Z3 and Z4 are position constraint coefficients.

[0034] Preferred, updated Value, calculate the prey's nonlinear escape energy factor ,according to In the steps of executing a global exploration strategy or a local development strategy, the numerical values ​​are used. The formula for calculating the value is expressed as:

[0035] ;

[0036] ;

[0037] In the formula: These are the initial values ​​of the weights at the start of the iteration; The final value at the end of the weight iteration; t is the current iteration number; T is the maximum iteration number; m is... Random numbers between The energy factor for the prey's escape; for Random numbers.

[0038] Preferred, updated Value, calculate the prey's nonlinear escape energy factor ,according to In the steps of executing a global exploration strategy or a local development strategy, when At that time, the algorithm executes a local development strategy, using a back-learning crossover operator perturbation strategy to update the optimal individual position information of the population under soft and hard encirclement conditions, as expressed in the following expression:

[0039] ;

[0040] In the formula: For the population iterative optimal solution The inverse solution; for A random number between [a certain number of points].

[0041] Preferably, when At that time, a soft encirclement strategy is implemented to update the individual Harris Eagle positions; when At that time, the algorithm executes a progressive and rapid dive soft encirclement strategy to update the individual position, and uses the optimal solution under different strategies to perform reverse learning perturbation to update the final optimal position of the individual in the population.

[0042] when When, execute the hard encirclement strategy and update the individual's position; when At that time, a gradual and rapid dive hard encirclement strategy is implemented to update the individual position, and the optimal solution under different strategies is used for reverse learning perturbation to update the final optimal position of the individual in the population.

[0043] The present invention provides an agricultural UAV trajectory planning method based on the improved Harris Eagle algorithm, which combines Chebyshev mapping, nonlinear escape energy factor mechanism and reverse learning cross operator perturbation strategy. By constructing a trajectory planning model that integrates constraints such as UAV flight area and maximum climb angle, a safe, reliable and low-cost UAV trajectory planning scheme is obtained. Attached Figure Description

[0044] Figure 1 A schematic diagram of an environment model provided for an embodiment of the present invention;

[0045] Figure 2 A flowchart of an agricultural UAV trajectory planning method based on an improved Harris Eagle algorithm provided in an embodiment of the present invention;

[0046] Figure 3 The convergence curve of the test function provided in the embodiment of the present invention;

[0047] Figure 4 Aerial topographic map of the actual scene experiment provided in the embodiments of the present invention;

[0048] Figure 5 A front view of an environmental model with a 20m radius area provided in an embodiment of the present invention;

[0049] Figure 6 A top view of an environmental model with a 20m radius provided in an embodiment of the present invention;

[0050] Figure 7 A front view of an environmental model with a 30m radius area provided in an embodiment of the present invention;

[0051] Figure 8 A top view of an environmental model with a radius of 30m provided in an embodiment of the present invention;

[0052] Figure 9A front view of an environmental model with a 40m radius area provided in an embodiment of the present invention;

[0053] Figure 10 This is a top view of the 40m radius area environment model provided in an embodiment of the present invention. Detailed Implementation

[0054] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0055] It is understood that the terms "first," "second," etc., used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish one element from another. For example, without departing from the scope of this application, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0056] like Figure 2 The diagram shows a flowchart of an agricultural UAV trajectory planning method based on an improved Harris Eagle algorithm provided in an embodiment of the present invention. The method includes:

[0057] Obtain the environmental conditions of the site, construct a site environment model, and determine the objective function based on the site environment model.

[0058] In this step, such as Figure 1 As shown, the trajectory planning of this invention is based on a specific on-site orchard environment model. This study uses a high-peak model to simulate obstacle areas, establishing a mathematical model as Equation (1):

[0059] (1);

[0060] In the formula, (x i y i Let h be the center coordinates of the i-th peak; i For terrain parameters, control the height; x si and y si , representing the attenuation along the x-axis and y-axis for the i-th peak, respectively, controlling the slope; n is the total number of peaks. The environmental model in this paper is as follows: Figure 1 As shown.

[0061] Terrain constraints: To avoid collisions during the actual flight of the drone in the orchard, this paper sets the flight altitude to be higher than the terrain altitude. Therefore, the model of the terrain constraint is Equation (2):

[0062] (2);

[0063] In the formula, Z i Z represents the drone's flight altitude, Z is the terrain function, and Z(x) i ,y i ) represents (x i ,y i The fitness level will be amplified proportionally when the flight altitude is less than or equal to the terrain altitude.

[0064] Position constraints: To prevent the UAV from colliding in unknown environments, the model for its position constraints is established as Equation (3):

[0065] (3);

[0066] In the formula, x max y max z max These represent the maximum values ​​in each dimension of the solution space. When a path point exceeds the solution space, the fitness will be amplified.

[0067] Maximum climb angle constraint: The UAV is also subject to the maximum climb angle constraint. If the maximum climb angle is exceeded, the UAV will stall and cause danger. Therefore, the climb angle γ constraint condition can be expressed as equation (4):

[0068] (4);

[0069] In this invention, when the flight path satisfies all constraints, the cost is the flight path length; conversely, when certain constraints are not met, the fitness of the position point is appropriately amplified through the control variables of the corresponding constraints. The specific expressions of the control variables of each constraint are Equations (5), (6), and (7):

[0070] (5);

[0071] (6);

[0072] (7);

[0073] Where c1, c2, and c3 are the control variables for terrain constraints, position constraints, and angle constraints, respectively, and P is the solution space of the optimization problem. The objective function expression based on each constraint condition is Equation (8):

[0074] (8);

[0075] Among existing technologies, the Harris Eagle Algorithm (HHO) is a novel metaheuristic algorithm. The algorithm's concept originates from the unique cooperative foraging behavior of Harris eagle populations.

[0076] Exploration Phase (Strategy 1): During the global exploration phase, the Harris Eagle randomly perches in the search space and observes prey. At a given location, the Harris Eagle will determine its target based on two equally probable factors. The size of the Harris Hawk is used to select an appropriate strategy to explore prey. The formula for updating the position vector of the Harris Hawk in the next iteration is Equation (9):

[0077] (9);

[0078] (10);

[0079] In the formula: An individual randomly selected from the current population; and For the current Harris Hawks The second iteration and the first The position vector of the next iteration; All Random numbers; This represents the current optimal position for the individual, i.e., the prey. The average position of individuals in the population; This represents the average positional difference between prey and individuals. and These are the lower and upper bounds of the population, respectively. This refers to the population size.

[0080] From global exploration to local exploitation: The HHO algorithm is based on the prey's escape energy factor. The size enables the transition from the global exploration phase to the local development phase. When When the prey escapes, the HHO algorithm enters the global exploration phase; otherwise, it enters the local exploration phase. During the prey's escape, its escape energy factor... The expression is given by equation (11):

[0081] (11);

[0082] In the formula: The energy factor for the prey's escape; for Random numbers; This represents the current iteration number; This represents the maximum number of iterations.

[0083] Partial development phase: based on the absolute value of escape energy and random numbers Based on comparisons with 0.5, and considering the hunting methods of Harris eagles and the escape behavior of their prey, four strategies can be identified: soft encirclement, hard encirclement, soft encirclement with gradual rapid dive, and hard encirclement with gradual rapid dive.

[0084] Soft encirclement (strategy 2): When When the prey has enough energy and a chance to escape, the Harris Eagle will hover around the prey to deplete its energy, employing a soft encirclement strategy to ambush it. The position update formula is Equation (12):

[0085] (12);

[0086] in:

[0087] (13);

[0088] (14);

[0089] In the formula: This is the difference between the optimal individual and the current individual; for Jump distance during the escape process; for Random numbers.

[0090] Hard encirclement (strategy 3): when At this point, the prey has neither enough energy to escape nor the opportunity to escape, so the Harris Eagle uses a hard encirclement strategy to ambush the prey. The position update formula is Equation (15):

[0091] (15);

[0092] Gradual, rapid, soft encirclement (strategy 4): When When the prey has enough energy to escape and has a chance to successfully escape from the encirclement, the Harris Eagle needs to form a soft encirclement before attacking, and then use a gradual and rapid dive strategy to capture the prey. The position is updated according to Equation (16) and compared with the fitness value of the current position. If the fitness value is improved, Equation (16) is used to update the position; otherwise, it means that the encirclement has failed, and Equation (17) is used to update the position.

[0093] (16);

[0094] (17);

[0095] In the formula: The dimensions of the problem; For one A dimensional random vector; The fitness function; For Levy flight functions.

[0096] At this stage, the final encirclement strategy is formula (18):

[0097] (18);

[0098] Gradual, rapid dive and hard encirclement (Strategy 5): When At this point, the prey has a chance to escape, but lacks the energy to do so. Therefore, the Harris Eagle first forms a tight encirclement before attacking, then employs a gradual, rapid dive strategy to surround and capture the prey. The position update formula is as follows:

[0099] (19);

[0100] (20);

[0101] (twenty one);

[0102] If the fitness value is updated, it is based on the average position. Update and execute the policy. If the fitness value is not updated, then the strategy is adopted. If the strategy fails, return to the original position.

[0103] In this invention, an improved Harris Hawk algorithm (CSHHO) is adopted. First, the Chebyshev mapping is introduced to complete the population initialization, which increases the diversity of the initial population while expanding the global exploration space, thereby improving the convergence speed and optimization range of the algorithm. Then, the escape energy factor is changed by using nonlinear logarithmic inertia weights, thereby balancing the global exploration and local exploitation capabilities of the algorithm. Finally, the ability of the best individual in the algorithm to escape local optima is improved by using a back-learning crossover operator perturbation strategy.

[0104] Population initialization using the Chebyshev map: Chaos is a deterministic, stochastic nonlinear dynamic system, and chaotic sequences are characterized by irregularity, randomness, and ergodicity. Since the last century, chaotic maps have received widespread attention in the field of optimization algorithms, as their dynamic stochastic properties enable optimization algorithms to perform more effective global searches within the search region.

[0105] Improved Harris Hawk algorithm for agricultural UAV trajectory planning: Although the standard HHO algorithm has a good convergence speed, when solving the global optimum problem, the randomly generated initial population may have poor population diversity, affecting the global search capability of the algorithm, resulting in slow convergence speed or unsatisfactory optimization accuracy. In order to improve the search efficiency of the algorithm, chaotic mapping is introduced into the population initialization stage of the algorithm. Through chaotic mapping, the initial individuals are evenly distributed in the search area, thereby improving the diversity of the initial population. In terms of the choice of chaotic mapping model, this invention chooses to introduce the population initialization operation of Chebyshev mapping. The traditional Chebyshev mapping can make it enter a chaotic state by controlling relevant parameters. Compared with one-dimensional mappings such as Tent mapping and Logistic mapping, it has a larger value range and good chaotic characteristics. The expression for the chaotic sequence generated by Chebyshev mapping is Equation (22):

[0106] (twenty two);

[0107] In the formula: This represents the order of chaos.

[0108] By setting the initial position of the Harris Eagle population using the Chebyshev mapping, equation (9) is updated to equation (23):

[0109] (twenty three);

[0110] The escape energy factor mechanism of nonlinear changes: In the HHO algorithm, whether the exploration behavior is successfully transformed into the exploit behavior depends on the escape energy factor. The absolute value of the number of iterations. hour, The number of iterations decreases linearly from 2 to 1, at which point the algorithm executes a global exploration strategy; when the number of iterations decreases from 2 to 1, the algorithm executes a global exploration strategy. hour, The energy decreases linearly from 1 to 0, and the algorithm executes a local exploitation strategy. However, in nature, the process of the Harris Eagle hunting its prey cannot be simulated by a single linear decreasing function. This linear escape energy reduction strategy easily leads to a poor balance between the global exploration and local exploitation behaviors of the algorithm. The algorithm is prone to getting stuck in local optima in the later stages of iteration, resulting in premature convergence. Inspired by the inertia weight setting of the PSO algorithm, this paper uses a nonlinear logarithmic inertia weight in the energy reduction formula to change the iterative process of escape energy, thereby balancing the global exploration and local exploitation capabilities of the algorithm. The expression is updated to equation (24):

[0111] (twenty four);

[0112] (25);

[0113] In the formula: These are the initial values ​​of the weights at the start of the iteration; The final value at the end of the weight iteration; t is the current iteration number; T is the maximum iteration number; m is... A random number between [a certain number of points].

[0114] Reverse learning perturbation strategy: In the later stages of the HHO algorithm iteration, relying solely on the guidance of the best individual in the population for population position updates leads to a high dependence on the globally optimal individual. When the optimal individual lacks the ability to escape local optima, the algorithm is prone to getting trapped in local extrema. Therefore, in the CSHHO algorithm, a reverse learning strategy is first used to capture the reverse solution in the solution space. Then, a crossover operator strategy is introduced to perturb the current optimal solution and the reverse solution, expanding the algorithm's global search space and enhancing population diversity, thereby improving the algorithm's ability to escape local optima.

[0115] The goal of reverse learning is to find the corresponding reverse solution in the search space based on the current solution, and then select and save the better solution through comparison. The mathematical model expression is given by equation (26):

[0116] (26);

[0117] In the formula: For the population iterative optimal solution The inverse solution; for A random number between [a certain number of points].

[0118] Logical flow:

[0119] Step 1: Initialize parameters, including population size Maximum number of iterations Dimensions of the objective function Initial value upper and lower bounds and .

[0120] Step 2: Initialize the population using Chebyshev mapping. Initialize individuals according to equation (23) and retain Harris Eagle individuals with higher fitness to form the final initial population.

[0121] Step 3: Update The value is calculated based on equation (24) to determine the prey's nonlinear escape energy factor. This enables the transition between global exploration and local development.

[0122] Step 4: When At that time, the algorithm executes a global exploration strategy and updates the individual position according to equation (9).

[0123] Step 5: When At this time, the algorithm executes a local development strategy. To avoid the algorithm getting trapped in local optima, a back-learning crossover operator perturbation strategy is used to update the optimal individual position information of the population under soft and hard encirclement conditions:

[0124] (1) When When the algorithm executes a soft encirclement strategy, it updates the individual Harris Eagle positions according to equation (12); when At that time, the algorithm executes a progressive rapid dive soft encirclement strategy, updates the individual position according to Equation (18), and uses Equation (26) to perform reverse learning perturbation on the optimal solution under the two strategies, and updates the final optimal position of the individual in the population.

[0125] (2) When When the algorithm executes a hard encirclement strategy, it updates the individual position according to equation (15); when At that time, the algorithm executes a progressive rapid dive hard encirclement strategy, updates the individual position according to Equation (19), and uses Equation (26) to perform reverse learning perturbation on the optimal solution under the two strategies, and updates the final optimal position of the individual in the population.

[0126] Step 6: By comparing the fitness values ​​of all individuals in the entire search space, continuously update the current global optimum and determine whether the maximum number of iterations has been reached. If the conditions are met, proceed to step 7; otherwise, proceed to step 2.

[0127] Step 7: Output the optimal individual fitness value and position.

[0128] The objective function is solved by improving the Harris Eagle algorithm:

[0129] Initialization parameters, including population size Maximum number of iterations Dimensions of the objective function Initial value upper and lower bounds and .

[0130] The population was initialized using Chebyshev mapping, and individuals were initialized. Individuals with higher fitness were retained to form the final initial population.

[0131] renew Value, calculate the prey's nonlinear escape energy factor ,according to The numerical values ​​can be used to execute a global exploration strategy or a local development strategy.

[0132] By comparing the fitness values ​​of all individuals in the entire search space, the current global optimal solution is continuously updated, and it is determined whether the maximum number of iterations T has been reached. If the condition is met, the fitness value and position of the optimal individual are output; otherwise, the population is reinitialized.

[0133] The effectiveness of this invention is demonstrated through the following experiments:

[0134] To test the optimization capabilities of the improved algorithms, this paper selects five benchmark functions to test them, and sets up comparisons between the EO, GWO, PSO, HHO, and CSHHO algorithms. The relevant attribute parameters of the functions are shown in Table 1. The initial population size of all algorithms is set to 30, and the number of iterations is set to 500.

[0135] Table 1 Benchmark Function Table

[0136]

[0137] The experiments statistically analyzed the optimal value, average value, and variance of each algorithm to measure their optimization ability. The results for the CSHHO algorithm and the comparison algorithms are shown in Table 2. The convergence curves are shown in the figure below. Figure 3 As shown in the figure, the horizontal axis represents the number of iterations, and the vertical axis represents the optimal fitness value. The comparison of algorithm test results shows that the CSHHO algorithm has certain advantages compared to the control algorithms. Under test functions F1, F4, and F5, the PSO algorithm outperforms the EO, GWO, and HHO algorithms. However, under test functions F2 and F3, the average and variance values ​​of the PSO algorithm are lower than those of the EO, GWO, and HHO algorithms, indicating that the PSO algorithm has the ability to escape local optima, but its optimization robustness is not strong. The CSHHO algorithm's results are all better than the results of the previous four algorithms, indicating that the CSHHO algorithm effectively improves the convergence speed, reduces the probability of getting trapped in local optima, and has a higher global search capability compared to the control algorithms. The convergence curves of the test functions show that, compared to the control algorithms, under the five test functions, the CSHHO algorithm has a shorter convergence time as the number of iterations increases. In summary, the CSHHO algorithm performs best compared to the other five algorithms.

[0138] Table 2 Comparison of Algorithm Test Results

[0139]

[0140] Field Experiment and Results: Field photography was conducted at the mulberry plantation in the Zhanjiang South Subtropical Botanical Garden, Mazhang District, Zhanjiang City, Guangdong Province. First, a drone was used to take aerial photographs of the overall terrain. Then, the experimental area was demarcated as follows: Figure 4 As shown.

[0141] Comparison of Algorithms Based on Obstacles of Different Radius: To increase the flight difficulty of the improved algorithm in UAV trajectory planning, this paper generates obstacle regions with radii of 20, 30, and 40m in the constructed environment model map based on real-world scene maps. Larger radii indicate greater path planning difficulty. Experiments were conducted in a Matlab 2018a, Windows 10, 16GB environment, with the flight start point set to (45, 45, 0), the target point to (200, 200, 80), a population size of 70, and 50 iterations. Simulation experiments were performed comparing the GWO, PSO, and HHO algorithms. The comparison results are shown in the figure below. Figure 5 , Figure 6 , Figure 7 , Figure 8 , Figure 9 and Figure 10 As shown, under constraints such as the maximum turning angle, the CSHHO algorithm can plan the shortest, i.e., optimal path between the starting and ending points of an obstacle-avoiding region in obstacle environment models with different radii, and can converge to the optimal trajectory.

[0142] This invention addresses the slow convergence speed and susceptibility to local optima of the HHO algorithm by proposing an improved CSHHO algorithm through a Chebyshev mapping strategy, a nonlinear escape energy factor mechanism, and a back-learning perturbation strategy. Benchmark tests verify that the CSHHO algorithm outperforms the compared algorithms. For the trajectory planning problem of agricultural UAVs in complex hilly orchard terrain, a safe trajectory planning model for agricultural UAVs is constructed based on constraints such as terrain, location, and maximum turning angle within an environmental model. The CSHHO algorithm is applied to solve the trajectory planning model. Comparative experiments on trajectory planning in environments with obstacles of different radii demonstrate that the CSHHO algorithm outperforms the GWO, PSO, and HHO algorithms. Experiments verify the effectiveness of the proposed safe UAV trajectory planning model method in complex hilly orchard terrain environments.

[0143] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0144] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0145] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for flight path planning of agricultural unmanned aerial vehicles based on an improved Harris Eagle algorithm, characterized in that, The method includes: Obtain the environmental conditions of the site, construct a site environment model, and determine the objective function based on the site environment model; The objective function is solved by improving the Harris Eagle algorithm: Initialization parameters, including population size Maximum number of iterations Dimensions of the objective function Initial value upper and lower bounds and ; The population was initialized using Chebyshev mapping, and individuals were initialized. Individuals with higher fitness were retained to form the final initial population. renew Value, calculate the prey's nonlinear escape energy factor ,according to The numerical values ​​can be used to execute a global exploration strategy or a local development strategy. By comparing the fitness values ​​of all individuals in the entire search space, the current global optimal solution is continuously updated, and it is determined whether the maximum number of iterations T has been reached. If the condition is met, the fitness value and position of the optimal individual are output; otherwise, the population is reinitialized. In the steps of obtaining the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, the mathematical model of the site environment model is expressed as follows: ； In the formula, (x i y i Let h be the center coordinates of the i-th peak; i For terrain parameters, control the height, x si and y si These represent the attenuation amounts of the i-th peak along the x-axis and y-axis, respectively, controlling the slope, and n is the total number of peaks; In the steps of acquiring the site's environmental conditions, constructing a site environment model, and determining the objective function based on the site environment model, terrain constraints are set, as follows: ； In the formula, Z i Z represents the drone's flight altitude, Z is the terrain function, and Z(x) i ,y i ) represents (x i ,y i The altitude of the terrain will be used to determine the fitness level. When the flight altitude is less than or equal to the terrain altitude, the fitness level will be amplified by a preset ratio. In the steps of acquiring the site's environmental conditions, constructing a site environment model, and determining the objective function based on the site environment model, location constraints are set, represented as follows: ； In the formula, x max y max and Z max These are the maximum values ​​in each dimension of the solution space, where n is a natural number. When a path point exceeds the solution space, the fitness will be amplified. In the steps of acquiring the environmental conditions of the site, constructing a site environment model, and determining the objective function based on the site environment model, the maximum climb angle constraint γ is set, expressed as: ； in, Maximum climb angle; The expression for the objective function is: ； In the formula: ; ； ； Where c1, c2, and c3 are the control variables for terrain constraints, position constraints, and angle constraints θ, respectively. max Let P be the maximum constraint angle, and let P be the solution space of the optimization problem.

2. The agricultural UAV trajectory planning method based on the improved Harris Eagle algorithm according to claim 1, characterized in that, In the step of initializing the population using Chebyshev mapping, where individuals are initialized and those with higher fitness are retained to form the final initial population, the initial position of the Harris Eagle population is set using Chebyshev mapping. Therefore, the initial population position is: ； in, An individual randomly selected from the current population; and For the current Harris Hawks The second iteration and the first The position vector of the next iteration; This represents the current optimal position for the individual, i.e., the prey. The average position of an individual in the population; The average positional difference between prey and individuals; and These are the lower and upper bounds of the population, respectively. For population size; r1, r 2、 r3 and r4 are position constraint coefficients.

3. The agricultural UAV trajectory planning method based on the improved Harris Eagle algorithm according to claim 2, characterized in that, renew Value, calculate the prey's nonlinear escape energy factor ,according to In the steps of executing a global exploration strategy or a local development strategy, the numerical values ​​are used. The formula for calculating the value is expressed as: ； ； In the formula: These are the initial values ​​of the weights at the start of the iteration; This is the termination value at the end of the weight iteration; t is the current iteration number; T is the maximum number of iterations; m is... Random numbers between The energy factor for the prey's escape; for Random numbers.

4. The agricultural UAV trajectory planning method based on the improved Harris Eagle algorithm according to claim 3, characterized in that, renew Value, calculate the prey's nonlinear escape energy factor ,according to In the steps of executing a global exploration strategy or a local development strategy, when At that time, the algorithm executes a local development strategy, using a back-learning crossover operator perturbation strategy to update the optimal individual position information of the population under soft and hard encirclement conditions, as expressed in the following expression: ； In the formula: For the population iterative optimal solution The inverse solution; for A random number between [a certain number of points].

5. The agricultural UAV trajectory planning method based on the improved Harris Eagle algorithm according to claim 4, characterized in that, when At that time, a soft encirclement strategy is implemented to update the individual Harris Eagle positions; when At that time, the algorithm executes a progressive and rapid dive soft encirclement strategy to update the individual position, and uses the optimal solution under different strategies to perform reverse learning perturbation to update the final optimal position of the individual in the population. when When, execute the hard encirclement strategy and update the individual's position; when At that time, a gradual and rapid dive hard encirclement strategy is implemented to update the individual position, and the optimal solution under different strategies is used for reverse learning perturbation to update the final optimal position of the individual in the population.

Images