A three-dimensional route planning method for fixed-wing UAV in complex mountainous areas

Abstract

The present application relates to the technical field of UAV path planning, and specifically discloses a three-dimensional route planning method for fixed-wing UAVs in complex mountains, which obtains mountain model information, determines the coordinates of the starting point and the end point, and generates a complex mountain map; by setting constraints on the safe turning radius and pitch angle of the fixed-wing UAV, a safer and more efficient flight of the fixed-wing UAV is achieved; at the same time, by introducing the golden sine strategy, Bernoulli chaos mapping, dynamic adaptive weight strategy and adaptive Gauss-Cauchy mixed variation perturbation strategy, the algorithm's regional development and global search capabilities and convergence performance are improved. By comparing the algorithm performance of the improved multi-strategy dung beetle optimization algorithm with the basic dung beetle optimization algorithm on 8 benchmark test functions, the results show that the improved multi-strategy dung beetle optimization algorithm has faster convergence speed and stability. The purpose of this patent is to solve the problems of insufficient global search capabilities and slow convergence speed in existing UAV route planning methods.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle (UAV) path planning, and in particular to a three-dimensional route planning method for a fixed-wing UAV in complex mountainous areas. Background Art

[0002] With the rapid development of science and technology, fixed-wing UAVs are widely used in many fields such as agriculture, surveying and mapping, environmental monitoring, emergency response and infrastructure inspection. Route planning is one of the core issues in the application of fixed-wing UAVs. The flight parameter constraints and route planning algorithms of fixed-wing UAVs are the two major factors affecting the final route planning. Therefore, constraining the flight parameters of fixed-wing UAVs and improving the route planning algorithm will help improve the safety, stability and flight efficiency of fixed-wing UAVs.

[0003] In terms of flight parameter constraints of fixed-wing UAVs, they are traditionally simplified to a mass point model. However, due to the large size of fixed-wing UAVs, they may collide with obstacles during flight. In terms of route planning algorithms, although traditional algorithms such as Dijkstra algorithm, artificial potential field method (APF) and simulated annealing algorithm (SA) are commonly used in this field, their convergence speed is slow. In recent years, intelligent optimization algorithms, such as A algorithm, genetic algorithm (GA), ant colony optimization algorithm (ACO) and particle swarm optimization algorithm (PSO), have been widely used in UAV route planning. A large number of studies have been made based on the above algorithms to improve them. For example, an improved A algorithm reduces the amount of calculation and memory overhead by expanding through jump point screening, thereby improving planning efficiency in large-scale scenarios. A route planning method for plant protection UAVs based on genetic algorithm, which constructs a super network by converting the operation route and scheduling route into virtual nodes or arcs, and verifies its feasibility in route planning for multiple plots of pesticide spraying. An improved particle swarm algorithm improves the optimality and real-time performance of the algorithm by combining step-by-step inertia weight adjustment and jumping out of the local optimal strategy. Although these algorithms have improved the route planning effect to a certain extent, insufficient global search ability and slow convergence speed are still common problems in existing route planning algorithms. Summary of the invention

[0004] In view of the deficiencies in the prior art, the technical problem solved by the present invention is to provide a three-dimensional route planning method for fixed-wing UAVs in complex mountainous areas, thereby solving the problems of insufficient global search capability and slow convergence speed in the existing UAV route planning methods.

[0005] In order to solve the above problems, the technical solution adopted by the present invention is: a three-dimensional route planning method for a fixed-wing UAV in complex mountainous areas, comprising the following steps:

[0006] S100: Acquire mountain model information, determine the coordinates of the starting point and the end point, and generate a complex mountain map;

[0007] S200: Construct the objective function with safe turning radius and pitch angle as constraints, and set the maximum number of iterations and dung beetle population size ,

[0008] S300: Initialize the position of the dung beetle using Bernoulli chaotic mapping;

[0009] S400: Calculate the routes of all individuals within a set range according to the objective function;

[0010] S500: updating the positions of all dung beetles by using the golden sine algorithm, checking whether the dung beetles are beyond the boundary and updating the positions of the dung beetles beyond the boundary of the map until the positions of all dung beetles are within the boundary;

[0011] S600: Introduce the global optimal solution of the previous generation, so that the current position is affected by the previous generation position and the optimal solution at the same time; update the current optimal solution and its route;

[0012] S700: Adopt the adaptive Gauss-Cauchy hybrid mutation perturbation strategy to perturb the current optimal solution and generate a new solution;

[0013] S800: Determine whether the set number of iterations has been reached ; If yes, stop iteration and return to the optimal route; if not, return to step S400 to continue iteration to find the optimal route.

[0014] The technical principle and beneficial effects of this scheme are as follows: by setting constraints on the safe turning radius and pitch angle of fixed-wing UAVs, the fixed-wing UAVs can be made to fly more safely and efficiently; at the same time, the Bernoulli chaotic mapping strategy is used to improve the diversity of the initial solution of the population, the golden sine strategy is used to enhance the convergence performance of the algorithm, the dynamic adaptive weight strategy is introduced to reduce the risk of falling into the local optimum, and the adaptive Gauss-Cauchy mixed mutation perturbation strategy is used to help the algorithm jump out of the local optimum and finally find the optimal solution in the world. The algorithm's regional development and global search capabilities and convergence performance are significantly improved. By comparing the algorithm performance of the improved multi-strategy dung beetle optimization algorithm of this scheme with the basic dung beetle optimization algorithm on 8 benchmark test functions, the results show that the improved multi-strategy dung beetle optimization algorithm has a faster convergence speed and stability. Through the route planning simulation experiment, the results show that the improved multi-strategy dung beetle optimization algorithm has better accuracy and convergence performance in the route planning of fixed-wing UAVs.

[0015] Furthermore, in step S300, firstly, a chaotic variable value is generated by Bernoulli chaotic mapping, and then it is mapped back to the initial space of the basic DBO algorithm by linear transformation; the specific expression of Bernoulli mapping is:

[0016]

[0017] In the formula Refers to the mapping parameters, .

[0018] The original dung beetle algorithm, namely the DBO algorithm, starts by randomly initializing the population position, but it is easy to fall into the local optimum due to the uneven distribution of individual positions, insufficient overall exploration ability and limited population diversity. This scheme uses chaotic mapping to generate a highly diverse initial population to increase the diversity of population solutions. Chaotic mapping combines determinism and randomness. Its randomness and non-periodic characteristics make it an effective substitute for random variables in intelligent algorithms, thereby expanding the scope of the search space and enhancing the search breadth of the optimization algorithm.

[0019] Furthermore, in step S500, the golden sine algorithm is used to optimize the location update algorithm of the existing DBO algorithm, and the corrected location update formula is as follows:

[0020]

[0021] In the formula, is a random number that determines the individual moving distance in the next iteration, is a random number that determines the direction of movement, j=1,2,3,...,d, Q is a random number that follows a normal distribution, L is a 1×d matrix, and all element values ​​are 1. ST is a safety value, , , According to the golden ratio The obtained coefficients.

[0022] The improved position update formula significantly improves the algorithm performance in two aspects. First, in each iteration, all individuals will exchange information with the current optimal individual to maximize the use of the optimal information, solving the problem of insufficient information exchange between individuals in the original algorithm. Secondly, in view of the limitation that the original algorithm is prone to fall into the local optimum, the introduction of the golden sine race algorithm expands the search capability and enhances the global search performance. At the same time, the search space is further compressed through the golden section function, and the search distance and direction of the dung beetle are reasonably controlled to optimize the optimization method of the DBO algorithm, achieving a balance between global exploration and local development.

[0023] Furthermore, in step S600, a dynamic weight coefficient is added ,coefficient The position update formula of the global optimal solution introduced in the previous generation is as follows;

[0024]

[0025]

[0026] In the formula, For the previous generation Dimensional global optimal solution; is a random number; is the warning value; is a safe value.

[0027] In the early iteration stage of the existing DBO algorithm, the algorithm tends to over-focus on local areas, resulting in falling into local optimality. To alleviate this problem, this solution introduces the global optimal solution of the previous generation, so that the current position is affected by both the previous generation position and the optimal solution. In addition, a dynamic weight coefficient is added to enhance the global search capability in the early stage, and to accelerate the local search and convergence process by reducing adaptability in the later stage.

[0028] Further, in step S700, the adaptive Gaussian-Cauchy mixture disturbance variation formula is as follows:

[0029]

[0030] In the formula, , , For individuals in The optimal position in the iteration, For the The optimal position in the iteration is the position after the Gaussian-Cauchy mixture perturbation, is the Gaussian mutation operator, is the Cauchy mutation operator.

[0031] In the late iteration of the existing dung beetle algorithm, individuals quickly tend to assimilate, and the entire group gradually gathers at the current local optimal position and accelerates convergence, but this will lead to a decrease in population diversity, making it difficult to jump out of the local optimum, and even stagnation may occur. To overcome this problem, this scheme proposes an adaptive Gaussian-Cauchy mixed perturbation mutation strategy. Considering the uncertainty of the mutation result, it is decided to perturb only the optimal individual, and by comparing the positions before and after the mutation, make full use of its information, enhance population diversity, and expand the search range. During the iteration process, individuals are perturbed in a specific way. Since the population distribution is uneven in the early stage of the algorithm and the individual positions are quite different, the Cauchy distribution is used for large-scale perturbations to increase diversity, help the algorithm jump out of the local optimum and achieve global exploration. As the iteration proceeds, the individual position gradually tends to be stable, and the algorithm performs small perturbations through Gaussian distribution in the later stage to accelerate convergence and maintain robustness in a high-dimensional environment. In summary, the adaptive Gaussian-Cauchy mixed mutation strategy of this scheme helps to achieve an effective balance between global exploration and local development.

[0032] Furthermore, the objective function of the fixed-wing UAV model in step S200 is constructed as follows: Assume that the fixed-wing UAV passes through N nodes during operation, and the coordinates are (x1, y1, z1), (x2, y2, z2), ..., (x N ,y N ,z N );

[0033] S201: The total route length formula for fixed-wing UAVs is:

[0034]

[0035] In the formula, Indicates the total length of the entire route, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction, and N represents the total number of route nodes;

[0036] S202: The formula for the altitude of a fixed-wing UAV route is:

[0037]

[0038] In the formula, Indicates the route altitude. represents the Z coordinate of the ith waypoint, represents the mean of all Z coordinates, and N represents the total number of route nodes;

[0039] S203: The formula for the safe turning radius of fixed-wing UAV flight is:

[0040]

[0041] In the formula, Indicates the safe turning radius, represents the change in the X coordinate of the ith route segment, Indicates the change in the Y coordinate of the i-th route segment, represents the change in the Z coordinate of the ith route segment, Indicates the change in the X coordinate of the i+1th route segment, Indicates the change in the Y coordinate of the i+1th route segment, represents the change in the Z coordinate of the i+1th route segment, and N represents the total number of route nodes;

[0042] S204: The pitch angle formula for fixed-wing UAV flight is:

[0043]

[0044] In the formula, is the pitch angle, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction;

[0045] S205: Overall objective function expression:

[0046]

[0047] In the formula, is the weight of each objective function. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is the algorithm flow chart of this method.

[0049] Figure 2 This is the path planning effect without safety turning radius and pitch angle constraints.

[0050] Figure 3 This is the path planning effect diagram with safe turning radius and pitch angle constraints.

[0051] Figure 4 Convergence curves of algorithms for 9 benchmark functions.

[0052] Figure 5 Comparison of route planning effects of different algorithms on different maps.

[0053] Figure 6 The curves of the number of iterations and optimal values ​​of different algorithms.

[0054] Figure 7 Comparison of route planning effects of MDBO and its variants on different maps.

[0055] Figure 8 Iteration times and optimal value curves of MDBO and its variants. DETAILED DESCRIPTION

[0056] The following is further described in detail through specific implementation methods:

[0057] Basically as attached Figure 1 As shown:

[0058] A method for three-dimensional route planning of a fixed-wing UAV in complex mountainous areas comprises the following steps:

[0059] S100: Acquire mountain model information, determine the coordinates of the starting point and the end point, and generate a complex mountain map;

[0060] S200: Use safe turning radius and pitch angle as constraints to build the objective function and set the maximum number of iterations and dung beetle population size ;

[0061] The objective function of the fixed-wing UAV model is constructed as follows: Assume that the fixed-wing UAV passes through N nodes during operation, and the coordinates are (x1, y1, z1), (x2, y2, z2), ..., (x N ,y N ,z N );

[0062] S201: The total route length formula for fixed-wing UAVs is:

[0063]

[0064] In the formula, Indicates the total length of the entire route, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction, and N represents the total number of route nodes;

[0065] S202: The formula for the altitude of a fixed-wing UAV route is:

[0066]

[0067] In the formula, Indicates the route altitude. represents the Z coordinate of the ith waypoint, represents the mean of all Z coordinates, and N represents the total number of route nodes;

[0068] S203: The formula for the safe turning radius of fixed-wing UAV flight is:

[0069]

[0070] In the formula, Indicates the safe turning radius, represents the change in the X coordinate of the ith route segment, Indicates the change in the Y coordinate of the i-th route segment, represents the change in the Z coordinate of the ith route segment, Indicates the change in the X coordinate of the i+1th route segment, Indicates the change in the Y coordinate of the i+1th route segment, represents the change in the Z coordinate of the i+1th route segment, and N represents the total number of route nodes;

[0071] S204: The pitch angle formula for fixed-wing UAV flight is:

[0072]

[0073] In the formula, is the pitch angle, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction;

[0074] S205: Overall objective function expression:

[0075]

[0076] In the formula, is the weight of each objective function.

[0077] S300: Initialize the position of the dung beetle using the Bernoulli chaotic map; specifically, first, generate chaotic variable values ​​through the Bernoulli chaotic map, and then map them back to the initial space of the basic dung beetle algorithm through linear transformation; the specific expression of the Bernoulli map is:

[0078]

[0079] In the formula Refers to the mapping parameters, .

[0080] S400: Calculate the routes of all individuals within a set range according to the objective function;

[0081] S500: Update the positions of all dung beetles using the golden sine algorithm, check whether the dung beetles are beyond the boundary, and update the positions of the dung beetles that are beyond the map boundary; the position update formula is as follows:

[0082]

[0083] In the formula, is a random number that determines the individual moving distance in the next iteration, is a random number that determines the direction of movement. , According to the golden ratio The obtained coefficients.

[0084] S600: Introducing the global optimal solution of the previous generation, so that the current position is affected by both the previous generation position and the optimal solution; updating the current optimal solution and its route; and adding a dynamic weight coefficient ,coefficient The position update formula of the global optimal solution introduced in the previous generation is as follows;

[0085]

[0086]

[0087] In the formula, For the previous generation Dimensional global optimal solution; is a random number; is the warning value; is a safe value.

[0088] S700: Adopt the adaptive Gauss-Cauchy mixed perturbation strategy to perturb the current optimal solution and generate a new solution; the adaptive Gauss-Cauchy mixed perturbation mutation formula is as follows:

[0089]

[0090] In the formula, , , For individuals in The optimal position in the iteration, For the The optimal position in the iteration is the position after the Gaussian-Cauchy mixture perturbation, is the Gaussian mutation operator, is the Cauchy mutation operator.

[0091] S800: Determine whether the set number of iterations has been reached ; If yes, stop iteration and return to the optimal route; if not, return to step S400 to continue iteration to find the optimal route.

[0092] This solution aims at the route planning problem of fixed-wing UAVs in complex mountainous environments, and sets flight parameter constraints such as safe turning radius and pitch angle to improve the safety, efficiency and stability of fixed-wing UAVs. At the same time, this solution improves the multi-strategy dung beetle optimization algorithm, namely the MBDO algorithm. The MBDO algorithm of this solution improves the diversity of the initial solution of the population through the Bernoulli chaotic mapping strategy, adopts the golden sine strategy to enhance the convergence performance of the algorithm, introduces a dynamic adaptive weight strategy to reduce the risk of falling into the local optimum, and uses the adaptive Gauss-Cauchy mixed mutation perturbation strategy to help the algorithm jump out of the local optimum and finally find the optimal solution in the global.

[0093] The superiority of this scheme is verified by the following four experiments.

[0094] Experiment 1: Comparative experiment on the impact of constraint conditions on safety

[0095] In the route planning of fixed-wing UAVs, the safe turning radius and pitch angle are two key constraints. If these constraints are not taken into account, the fixed-wing UAV may perform overly aggressive turns and pitch movements during flight, which not only affects flight efficiency but also may cause safety risks. Especially in complex terrain, the lack of appropriate turning radius and pitch angle constraints will significantly increase the risk of collision. By introducing safe turning radius and pitch angle constraints, route planning is safer. These constraints ensure that fixed-wing UAVs retain sufficient space and angles when turning and ascending to avoid stalling and collision. Specifically, the safe turning radius limits the minimum curvature radius of the fixed-wing UAV to prevent stalling or structural damage caused by sharp turns; the pitch angle constraint controls the ascent and descent angles of the fixed-wing UAV to ensure lift and control stability during flight.

[0096] The route planning was performed using the objective function without constraints and the objective function with constraints of safe turning radius and pitch angle. The experimental results are shown in Figure 2 , Figure 3 Experiments show that after adding these constraints, the flight path of the fixed-wing UAV will be smoother, and the turning and lifting actions will be gentler, thereby improving the safety and reliability of the flight.

[0097] Experiment 2: Algorithm performance analysis experiment based on benchmark function

[0098] In order to test the performance of the MDBO algorithm improved by this scheme, this experiment uses the eight benchmark functions listed in Table 1 to verify and analyze the MDBO algorithm and compare it with the DBO algorithm. These functions include unimodal benchmark functions, multimodal benchmark functions, and fixed-dimensional multimodal benchmark functions. The population size of all algorithms is set to 30, and the maximum number of iterations is 500.

[0099] Table 1 Benchmark functions

[0100]

[0101] Table 2 Experimental results of benchmark functions

[0102]

[0103] Each algorithm was run independently 30 times on 8 benchmark functions, and the optimal value, average value, and standard deviation of the function were recorded. The calculation results are shown in Table 2. As can be seen from Table 2, based on the comprehensive experimental results of 8 benchmark functions, the MDBO algorithm is consistent with the DBO algorithm in the optimal values ​​of the 7th and 8th functions. In the remaining 6 functions, the optimization accuracy of the MDBO algorithm is better than that of the DBO algorithm. At the same time, from the perspective of computational stability, the standard deviation of the MDBO algorithm in 30 runs is smaller than that of the DBO algorithm, indicating that the MDBO algorithm has stronger computational consistency and better optimization ability than the DBO algorithm.

[0104] Figure 4 The convergence curves of each algorithm are shown. The red one is the DBO algorithm and the green one is the improved MBDO algorithm. It can be seen from the figure that the improved MDBO algorithm shows faster convergence speed on all test functions and has outstanding optimization ability in different test functions.

[0105] Experiment 3: Route planning simulation experiment

[0106] In order to verify the robustness and effectiveness of the improved algorithm, this experiment used MATLAB to conduct comparative experiments on MDBO with PSO, COA, NGO, WOA, SSA and DBO algorithms in three mountain environments with different complexities. The specific parameter settings of each algorithm are shown in Table 3.

[0107] Table 3 Parameter variable settings

[0108]

[0109] In order to obtain fair experimental results, the common parameters of all algorithms are set consistently, as follows: the starting point coordinates are (0, 0, 10), the end point coordinates are (200, 200, 30), the population size is 50, and the maximum number of iterations is 50.

[0110] In three mountain maps with different complexity, the route optimization of each algorithm is as follows: Figure 5 and Figure 6 As shown. Among them, Figure 5 The comparison of route planning effects is shown. Figure 6 The curve comparison between the number of iterations and the optimal value is shown. In the figure, magenta represents the PSO algorithm, black represents the COA algorithm, blue represents the NGO algorithm, yellow represents the WOA algorithm, navy represents the SSA algorithm, red represents the DBO algorithm, and green represents the improved MBDO algorithm of this solution.

[0111] from Figure 6 It can be seen from the iterative curves of the three maps that the improved MDBO algorithm has a faster convergence speed, can search quickly in the early stage, and tends to converge before the 8th iteration. The convergence value is better than other algorithms, and the convergence speed and accuracy are better than other methods. The golden sine strategy proposed this time improves the convergence ability of the algorithm, and the introduced dynamic adaptive weight strategy reduces the probability of the algorithm falling into the local optimal solution. Through the adaptive Gauss-Cauchy mixed perturbation mutation strategy, the algorithm can effectively jump out of the local optimal solution. In addition, it can be seen from the figure that the DBO algorithm falls into the local optimal solution many times during the iteration process, which further verifies that the improved MDBO algorithm of this scheme successfully overcomes this drawback.

[0112] Table 4 Evaluation indicators in different maps

[0113]

[0114] Since the simulation experiment is random, 30 experiments were conducted on each of the three maps, and the longest route, shortest route, average and standard deviation of the fixed-wing UAV in these 30 experiments were statistically calculated. The results are shown in Table 4. As can be seen from Table 4, the improved MDBO algorithm is superior to other algorithms in all evaluation indicators. It is particularly noteworthy that the standard deviation of the MDBO algorithm is the smallest, which shows that the improved MDBO algorithm in this scheme has better stability and advantages in route optimization.

[0115] Experiment 4: Ablation experiment

[0116] In order to verify the necessity of introducing four strategies in improving the MDBO algorithm, a series of ablation experiments are designed. As shown in Table 5, the DBO algorithm with Bernoulli chaotic mapping, golden sine, and dynamic adaptive weight strategy is recorded as DBO123, the DBO algorithm with Bernoulli chaotic mapping, golden sine, and adaptive Gaussian-Cauchy mixed mutation perturbation strategy is recorded as DBO124, and the DBO algorithm with golden sine, dynamic adaptive weight, and adaptive Gaussian-Cauchy mixed mutation perturbation strategy is recorded as DBO234.

[0117] Table 5 MDBO and its variants

[0118]

[0119] In three mountain environments with different complexity, MATLAB was used to conduct comparative experiments between MDBO and DBO123, DBO124 and DBO234. At the same time, the common parameters were set to be consistent, as follows: starting point coordinates (0,0,10), end point coordinates (200,200,30), population size is 50, and maximum number of iterations is 50.

[0120] The route optimization results of each algorithm in three mountain maps with different complexity are as follows: Figure 7 and Figure 8 shown. Figure 7 This is a comparison chart of route planning effects. Figure 8 The figure is a comparison chart of the number of iterations and the optimal value curve. 30 experiments were conducted on each of the three maps, and the longest route, shortest route, average value and standard deviation of the fixed-wing UAV in the 30 experiments were statistically calculated and shown in Table 6.

[0121] Table 6 Evaluation indicators in different maps

[0122]

[0123] The experimental results show that the complete MDBO algorithm performs best in all evaluation indicators, whether in finding the longest and shortest routes or in terms of stability and reliability of the mean and standard deviation. The iterative curve graph further verifies the fast convergence characteristics of MDBO on different test maps, indicating that the combined effect of these strategies significantly improves the algorithm's route planning ability in complex mountainous environments.

[0124] Each strategy complements MDBO. Bernoulli chaotic mapping provides a wide range of search starting points for the algorithm by increasing the diversity of the initial solutions of the population; the golden sine algorithm improves the convergence performance; the dynamic adaptive weight strategy reduces the risk of falling into the local optimum by balancing the needs of global and local searches; the adaptive Gauss-Cauchy mixed mutation perturbation strategy introduces random perturbations, enhances the algorithm's exploration ability, helps it jump out of the local optimum, and finally finds the optimal solution in the global search. The integration of these strategies makes the MDBO algorithm perform well in the three-dimensional route planning problem of fixed-wing UAVs.

[0125] The above is only an embodiment of the present invention, and the common knowledge such as the known specific structure and characteristics in the scheme is not described in detail here. It should be pointed out that for those skilled in the art, several deformations and improvements can be made without departing from the structure of the present invention, which should also be regarded as the protection scope of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the specification can be used to interpret the content of the claims.

Claims

1. A three-dimensional route planning method for a fixed-wing UAV in complex mountainous terrain, characterized by: The following steps are included S100: Acquire mountain model information, determine the coordinates of the starting point and the end point, and generate a complex mountain map; S200: Use safe turning radius and pitch angle as constraints to build the objective function and set the maximum number of iterations and dung beetle population size , the objective function of the fixed-wing UAV model is constructed as follows: Assume that the fixed-wing UAV passes through N nodes during operation, and the coordinates are (x1, y1, z1), (x2, y2, z2), ..., (x N ,y N ,z N ); S201: The total route length formula for fixed-wing UAVs is: In the formula, Indicates the total length of the entire route, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction, and N represents the total number of route nodes; S202: The formula for the altitude of a fixed-wing UAV route is: In the formula, Indicates the route altitude. represents the Z coordinate of the ith waypoint, represents the mean of all Z coordinates, and N represents the total number of route nodes; S203: The formula for the safe turning radius of fixed-wing UAV flight is: In the formula, Indicates the safe turning radius, Indicates the change in the X coordinate of the ith route segment, represents the change in the Y coordinate of the i-th route segment, represents the change in the Z coordinate of the ith route segment, Indicates the change in the X coordinate of the i+1th route segment, Indicates the change in the Y coordinate of the i+1th route segment, represents the change in the Z coordinate of the i+1th route segment, and N represents the total number of route nodes; S204: The pitch angle formula for fixed-wing UAV flight is: In the formula, is the pitch angle, is the distance between two route nodes in the x-axis direction, is the distance between two route nodes in the y-axis direction, is the distance between two route nodes in the z-axis direction; S205: Overall objective function expression: In the formula, is the weight of each objective function; S300: Initializing the position of the dung beetle using chaos mapping; S400: Calculate the routes of all individuals within a set range according to the objective function; S5 00: Update the positions of all dung beetles using the golden sine algorithm, check whether the dung beetles are beyond the map boundary and update the positions of those that are beyond the map boundary until all dung beetle positions are within the boundary; S600: Introduced the global optimal solution of the previous generation, so that the current position is affected by the previous generation position and the optimal solution at the same time; updated the current optimal solution and its route; added a dynamic weight coefficient ,coefficient The position update formula of the global optimal solution introduced in the previous generation is as follows; In the formula, For the previous generation Dimensional global optimal solution; is a random number; is the warning value; is a safe value; S700: Adopt the adaptive Gauss-Cauchy hybrid mutation perturbation strategy to perturb the current optimal solution and generate a new solution; S800: Determine whether the set number of iterations has been reached ; If yes, stop iteration and return to the optimal route; if not, return to step S400 to continue iteration to find the optimal route.

2. A three-dimensional route planning method for a fixed-wing UAV in complex mountainous terrain according to claim 1, characterized in that: In step S300, first, a chaotic variable value is generated by Bernoulli chaotic mapping, and then it is mapped back to the initial space of the algorithm by linear transformation; the specific expression of Bernoulli mapping is: In the formula Refers to the mapping parameters, .

3. The method for three-dimensional route planning of a fixed-wing UAV in complex mountainous terrain according to claim 1, characterized in that: In step S500, the golden sine algorithm is used to optimize the location update algorithm of the existing DBO algorithm, and the corrected location update formula is as follows: In the formula, is a random number that determines the individual moving distance in the next iteration, is a random number that determines the direction of movement. , According to the golden ratio The obtained coefficients.

4. The method for three-dimensional route planning of a fixed-wing UAV in complex mountainous terrain according to claim 1, characterized in that: In step S700, the adaptive Gaussian-Cauchy mixture perturbation variation formula is as follows: In the formula, , , For individuals in The optimal position in the iteration, For the The optimal position in the iteration is the position after the Gaussian-Cauchy mixture perturbation, is the Gaussian mutation operator, is the Cauchy mutation operator.

Images