Unmanned aerial vehicle trajectory planning method based on multi-strategy improved dung beetle optimization algorithm

Abstract

The invention provides an unmanned aerial vehicle trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm, and relates to the technical field of unmanned aerial vehicle path planning. Firstly, an obstacle model in an unmanned aerial vehicle three-dimensional environment is established, and unmanned aerial vehicle performance constraints are designed. A dung beetle population containing a plurality of dung beetle individuals is constructed, the population is initialized by fusing Tent mapping and a random reverse learning strategy, and initial dung beetle individual positions are generated. And the dung beetle individual position is updated through a multi-strategy improved dung beetle optimization algorithm. And constructing a target cost function, calculating the fitness value of each dung beetle individual, and updating the optimal position of the individual. And adopting spline interpolation to obtain a global path curve of the unmanned aerial vehicle. And repeatedly executing the process from population construction to path generation until a preset maximum number of iterations is reached, and finally outputting the optimal trajectory of the unmanned aerial vehicle. The multi-strategy improved dung beetle optimization algorithm is used for unmanned aerial vehicle trajectory planning, and the accuracy of unmanned aerial vehicle trajectory planning is improved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of unmanned aerial vehicle path planning, and more particularly to an unmanned aerial vehicle trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm. BACKGROUND

[0002] In the civil aspect, the unmanned aerial vehicle technology has been widely penetrated into agriculture, building surveying and mapping, transportation logistics and fire rescue, etc. Correspondingly, in military applications, unmanned aerial vehicles mainly perform tasks such as battlefield intelligence collection, reconnaissance and surveillance, precision strike and assault operations, etc. With the increasing complexity of the operating environment of unmanned aerial vehicles, higher requirements are put forward for the efficiency and reliability of their flight path planning. As one of the key technologies, the flight path planning algorithm has been continuously paid close attention to by the academic circle, and the related research needs to be further deepened. At present, the common flight path planning methods can be divided into two categories: traditional classic algorithms and modern intelligent algorithms. Traditional unmanned aerial vehicle trajectory planning algorithms include A*, Rapidly-exploring Random Tree (RRT) and Artificial Potential Field (APF), etc., which are simple in structure and easy to implement, but due to the limited heuristic ability, they are easy to fall into local optimum, and the convergence speed is usually slower than that of modern intelligent algorithms. Intelligent algorithms represented by Artificial Fish Swarm Algorithm (AFSA), Harris Hawk Optimization (HHO) and Dung Beetle Optimization Algorithm (DBO) are being applied to various complex planning tasks more and more due to their superior robustness.

[0003] Dung Beetle Optimization Algorithm (DBO) is a novel swarm intelligence optimization algorithm, which is inspired by a series of intelligent behaviors of dung beetles (commonly known as "dung beetle") in nature. It simulates the behaviors of rolling balls, dancing, foraging, stealing and breeding of dung beetles to find the optimal solution of complex optimization problems. Since its proposal, it has been widely applied and researched in the fields of engineering optimization, parameter identification, machine learning model parameter tuning, path planning, etc. However, the existing Dung Beetle Optimization Algorithm (DBO) has the problems of easy premature convergence and poor balance between global exploration and local exploitation, which leads to poor accuracy of unmanned aerial vehicle trajectory planning and is difficult to meet the planning requirements of high reliability and high efficiency trajectories in complex three-dimensional environments. SUMMARY

[0004] In order to solve the problem of poor accuracy of current unmanned aerial vehicle trajectory planning, the application proposes an unmanned aerial vehicle trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm to improve the accuracy of unmanned aerial vehicle trajectory planning.

[0005] In order to achieve the above technical effects, the technical solutions of the application are as follows: S1: establishing an obstacle model in the three-dimensional environment of the unmanned aerial vehicle and designing performance constraints of the unmanned aerial vehicle; S2: Construct a group of dung beetle individuals; adopt the fusion Tent mapping and random reverse learning strategy to initialize the dung beetle population and generate the initial position of the dung beetle individual; S3: Update the position of the dung beetle individual by adopting the multi-strategy improved dung beetle optimization algorithm; S4: According to the performance constraints of the unmanned aerial vehicle designed in step S1, construct a target cost function, calculate the fitness value of each dung beetle individual, and update the best position of the dung beetle individual; S5: Obtain the global path curve of the unmanned aerial vehicle by using the obstacle model in the three-dimensional environment of the unmanned aerial vehicle and the best position of the dung beetle individual, and adopting spline interpolation; S6: Repeat steps S2-S5 until the maximum number of iterations is reached, and output the final optimal trajectory of the unmanned aerial vehicle.

[0006] Further, the expression of the obstacle model in the three-dimensional environment of the unmanned aerial vehicle is:

[0007] In the formula, represents the height function of the hill, represents the coordinates of the two-dimensional space point, represents the center coordinates of the hill, h represents the height parameter of the hill, and represents the steepness parameter of the hill.

[0008] Further, the performance constraints of the unmanned aerial vehicle include: length cost constraint, height cost constraint, yaw angle constraint and pitch angle constraint; The expression of the length cost constraint is:

[0009]

[0010] In the formula, represents the distance between the i th path node of the unmanned aerial vehicle and the i+1 th path node, represents the position value of the i th path node of the unmanned aerial vehicle in the x axis, y axis and z axis, represents the position value of the i+1 th path node of the unmanned aerial vehicle in the x axis, y axis and z axis, represents the length cost constraint; The expression of the height cost constraint is:

[0011] In the formula, ​represents a minimum flight height of the UAV, represents a maximum flight height of the UAV, represents a flight height of the UAV at the waypoint i, represents a path height cost of the UAV; The expression of the flight deviation angle constraint is:

[0012]

[0013]

[0014] In the formula, represents a flight deviation angle constraint, represents a flight deviation angle, and represents a minimum flight deviation angle and a maximum flight deviation angle of the UAV during flight, represents a vector of the i-th path segment; The expression of the pitch angle constraint is:

[0015]

[0016] In the formula, represents a flight deviation angle constraint, represents a pitch angle, represents a maximum climb angle of the UAV.

[0017] Further, the individual scarab beetles in the scarab beetle population include: rolling ball scarab beetles, breeding scarab beetles, foraging scarab beetles, and thief scarab beetles.

[0018] Further, the initial scarab beetle population is generated by fusing Tent mapping and random reverse learning strategies, and the process of generating the initial scarab beetle individual position is: The original solution of the scarab beetle individual position is generated by fusing Tent mapping; the reverse solution of the scarab beetle individual position with random disturbance is generated based on the original scarab beetle individual position by using the random reverse learning strategy; the candidate scarab beetle population set C is constructed based on the original solution of the scarab beetle individual position and the reverse solution of the scarab beetle individual position; the fitness value F of each scarab beetle individual in the candidate scarab beetle population set C is calculated by iteration, and the n scarab beetle individuals with the optimal fitness value F are selected to form the initial scarab beetle individual position; the expression of the process is:

[0019]

[0020] In the formula, represents the original position of the scarab beetle population, denotes the solution obtained by using the Tent map; denotes the control parameter of the Tent map; denotes the inverse solution of and denote the upper and lower limits of the position of the dung beetle population respectively, denotes the random inverse solution, and rand denotes a random number between 0 and 1.

[0021] Further, the process of updating the position of the dung beetle individual is as follows: the position of the rolling dung beetle individual is updated by using the rolling dung beetle optimization algorithm; the position of the breeding dung beetle individual is updated by using the breeding dung beetle optimization algorithm; the position of the foraging dung beetle individual is updated in advance by using the foraging dung beetle optimization algorithm; the updated position of the foraging dung beetle individual is calculated and updated by using the subtraction average strategy; the position of the thief dung beetle individual is updated in advance by using the thief dung beetle optimization algorithm; the position of the thief dung beetle individual is updated by using the adaptive t-distribution perturbation strategy and the Cauchy variation.

[0022] Further, the position of the rolling dung beetle individual is updated by using the rolling dung beetle optimization algorithm; the rolling behavior of the rolling dung beetle includes an obstacle mode and a non-obstacle mode; In the obstacle mode, the expression for updating the position of the rolling dung beetle individual is:

[0023] In the non-obstacle mode, the expression for updating the position of the rolling dung beetle individual is:

[0024] In the formula, t denotes the current iteration number, denotes the position of the rolling dung beetle individual n at the tth iteration, denotes a natural number, k denotes a deflection coefficient, and b denotes a constant between, denotes the absolute distance between the current rolling dung beetle individual and the global worst rolling dung beetle individual, denotes the global worst position, denotes the difference between the position of the n th rolling dung beetle individual at the t th iteration and the position at the (t-1) th iteration; The position of the breeding dung beetle individual is updated by using the breeding dung beetle optimization algorithm; the area where the breeding dung beetle is located includes an oviposition area and a breeding area; In the oviposition area, the expression for updating the position of the breeding dung beetle individual is:

[0025] In the breeding area, the expression for updating the position of the breeding dung beetle individual is:

[0026] wherein, denotes the local optimal dung beetle individual, and wherein denotes the upper and lower bounds of the simulated oviposition area, denotes the maximum number of iterations, and wherein denotes the lower and upper bounds of the search space; denotes the position of the brood ball n at the tth iteration, and wherein denotes a random vector, D denotes the dimension of the optimization problem, and R denotes the boundary convergence factor.

[0027] The foraging dung beetle individual position is updated in advance using the foraging dung beetle optimization algorithm, and the expression of the foraging dung beetle individual in the optimal foraging area is:

[0028] When the foraging dung beetle individual forages, the expression for updating the position of the foraging dung beetle individual is:

[0029] wherein, denotes the global optimal solution, and wherein denotes the lower and upper bounds of the optimal foraging area; denotes the position of the dung beetle individual n at the tth iteration, denotes a random number subject to a normal distribution, denotes a random vector; The update position of the foraging dung beetle individual is calculated using the subtraction average strategy, and the expression is:

[0030] wherein, denotes the update position of the foraging dung beetle individual, denotes a vector of dimension m, denotes the position of the foraging dung beetle individual j; The position of the foraging dung beetle individual is updated according to the update position of the foraging dung beetle individual, and the expression is:

[0031] wherein, denotes the position of the foraging dung beetle individual after updating, denotes a weight coefficient, denotes a weight coefficient, denotes the upper bound of the optimal foraging area, denotes a vector of dimension m, lower limit of the optimal foraging zone, position of the foraging scarab individual j at iteration t, denotes the subtraction average operation rule, N denotes the total number of individuals of the scarab population; The position of the thief scarab individual is updated in advance using the thief scarab optimization algorithm, and the expression is as follows:

[0032] In the formula, denotes the optimal food source, g denotes a random vector subject to a normal distribution, and s denotes a constant; The position of the thief scarab individual is updated using an adaptive t-distribution disturbance strategy and Cauchy mutation, and the expression is as follows:

[0033] In the formula, denotes the position of the thief scarab individual after adaptive t-distribution disturbance, denotes the optimal position of the current thief scarab individual; denotes a random number generated by adaptive t-distribution; denotes a random number generated by Cauchy distribution, and denote the contribution coefficients of adaptive t-distribution and Cauchy distribution.

[0034] Further, the expression of the adaptive t-distribution disturbance strategy is as follows:

[0035] In the formula, m denotes a degree of freedom parameter, denotes a gamma function; The expression of the Cauchy mutation is as follows:

[0036] In the formula, denotes Cauchy mutation.

[0037] Further, the expression of the target cost function is as follows:

[0038]

[0039] In the formula, F denotes the target cost function, , , , denotes an evaluation weight.

[0040] Further, the process of obtaining the global path curve of the unmanned aerial vehicle by using the obstacle model in the three-dimensional environment of the unmanned aerial vehicle and the optimal position of the individual scarab and adopting spline interpolation is: taking the segmented intersection points of the spline interpolation as path nodes, and all the path nodes of a path form a set; the three-dimensional coordinates of the path nodes are split into three sets according to the dimensions; the three sets are merged to obtain the coordinates of the optimal position of the individual scarab; and the global path curve of the unmanned aerial vehicle is obtained based on the coordinates of the optimal position of the individual scarab and in combination with the set starting point and target point.

[0041] Compared with the prior art, the method has the beneficial effects that: The application provides an unmanned aerial vehicle trajectory planning method based on a multi-strategy improved scarab optimization algorithm, and relates to the technical field of unmanned aerial vehicle path planning. First, an obstacle model in a three-dimensional environment of an unmanned aerial vehicle is established, and performance constraints of the unmanned aerial vehicle are designed. A scarab population containing a plurality of individual scarabs is constructed, an initial position of the individual scarab is generated by initializing the population by using a fusion Tent mapping and a random reverse learning strategy, and the ability of global exploration and local development is considered. The position of the individual scarab is updated by using the multi-strategy improved scarab optimization algorithm, the search efficiency is improved, and premature convergence is avoided. A target cost function is constructed, the fitness value of each individual scarab is calculated, and the optimal position of the individual scarab is updated. Spline interpolation is adopted to obtain the global path curve of the unmanned aerial vehicle. The process of population construction to path generation is repeatedly executed until a preset maximum iteration number is reached, and finally the optimal trajectory of the unmanned aerial vehicle is output. The multi-strategy improved scarab optimization algorithm proposed in the application improves the accuracy of the trajectory planning of the unmanned aerial vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 A flowchart of the unmanned aerial vehicle trajectory planning method based on the multi-strategy improved scarab optimization algorithm proposed in the embodiment of the application is shown; Figure 2 A top view of the unmanned aerial vehicle trajectory planning result of the unmanned aerial vehicle trajectory planning method based on the multi-strategy improved scarab optimization algorithm proposed in the embodiment of the application is shown; Figure 3 A standard view of the unmanned aerial vehicle trajectory planning result of the unmanned aerial vehicle trajectory planning method based on the multi-strategy improved scarab optimization algorithm proposed in the embodiment of the application is shown. DETAILED DESCRIPTION

[0043] The drawings are only used for illustrative description, and cannot be understood as a limitation on the patent; In order to better illustrate the embodiment, some parts of the drawings are omitted, enlarged or reduced, and do not represent the actual size; For those skilled in the art, it is understandable that some well-known content in the drawings can be omitted.

[0044] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0045] The positional relationships depicted in the accompanying drawings are for illustrative purposes only and should not be construed as limiting this patent. Example 1 This embodiment proposes a UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm, such as... Figure 1 The flowchart of the method shown includes the following steps: S1: Establish an obstacle model in the 3D environment of the UAV and design UAV performance constraints; S2: Construct a dung beetle population containing several dung beetle individuals; use a fusion of Tent mapping and random back-learning strategy to initialize the dung beetle population and generate the initial dung beetle individual locations; S3: Employ a multi-strategy improvement algorithm for dung beetle optimization to update the individual dung beetle positions; S4: Based on the performance constraints of the drone designed in step S1, construct the objective cost function, calculate the fitness value of each dung beetle individual, and update the optimal position of the dung beetle individual; S5: Using the obstacle model in the UAV's 3D environment and the optimal position of the dung beetle, spline interpolation is used to obtain the global path curve of the UAV; S6: Repeat steps S2-S5 until the preset maximum number of iterations is reached, and output the final optimal trajectory of the UAV.

[0046] In this embodiment, the obstacles near the UAV must first be modeled. During the modeling process, the terrain is divided into plains and hills, with hills having the most significant impact. The expression for the obstacle model in the UAV's 3D environment is:

[0047] In the formula, The function representing the height of the hill. Represents the coordinates of a point in two-dimensional space. Indicates the center coordinates of the hill. h The parameter representing the height of the hill. and This represents the steepness parameter of the hill.

[0048] The performance constraints of unmanned aerial vehicles (UAVs) include: length cost constraints, altitude cost constraints, yaw angle constraints, and pitch angle constraints. The expression for the length cost constraint is as follows:

[0049]

[0050] In the formula, Represented as the i-th path node of the UAV With the (i+1)th path node The distance between them This represents the position values ​​of the i-th path node of the UAV on the x-axis, y-axis, and z-axis. This represents the position values ​​of the (i+1)th path node of the drone on the x, y, and z axes. Indicates length cost constraint; The expression for the high cost constraint is:

[0051] In the formula, Indicates the minimum flight altitude of the drone. Indicates the maximum flight altitude of the drone. This indicates the flight altitude of the drone at waypoint i. Indicates the cost of drone path altitude; Altitude cost constraints penalize flights that exceed their upper and lower limits to maintain an average altitude.

[0052] The maximum yaw angle of a UAV refers to the maximum angle of change in the projection of the UAV's flight path onto the horizontal plane. Due to its own performance and inertia, a UAV, if it needs to change its flight direction during flight, can only turn within the maximum yaw angle range and cannot fly at arbitrary angles. The expression for the yaw angle constraint is:

[0053]

[0054]

[0055] In the formula, This indicates the yaw angle constraint. Indicates the yaw angle. and This indicates the minimum and maximum yaw angles of the drone during flight. The vector representing the i-th path segment. and Let represent the coordinates of the i-th path node and the (i+1)-th path node, respectively; The maximum pitch angle of a drone refers to the maximum angle between the drone's axis and the horizontal plane during ascent or descent. Due to its own performance constraints, and for safety and speed control, the angle must be less than the maximum pitch angle when climbing or descending during flight. The expression for the pitch angle constraint is:

[0056]

[0057] wherein, represents the yaw angle constraint, represents the pitch angle, represents the maximum climb angle of the UAV.

[0058] Set the number of individual dung beetle in the dung beetle population as N; the upper and lower bounds of the dung beetle population as Ub and Lb; the maximum number of iterations as ; and randomly initialize the position of each dung beetle. The individual dung beetles in the dung beetle population include: rolling dung beetles, breeding dung beetles, foraging dung beetles and thief dung beetles.

[0059] The process of generating the initial position of the individual dung beetle is as follows by using the Tent mapping and the random reverse learning strategy: The original solution of the position of the individual dung beetle is generated by using the Tent mapping (having good ergodicity); the reverse solution of the position of the individual dung beetle with random disturbance is generated based on the original position of the individual dung beetle by using the random reverse learning strategy; the candidate dung beetle population set C (having a size of 2n) is constructed based on the original solution of the position of the individual dung beetle and the reverse solution of the position of the individual dung beetle; the fitness value F of each individual dung beetle in the candidate dung beetle population set C is calculated by iteration, and the n individual dung beetles with the optimal fitness value F are selected to form the initial position of the individual dung beetle. The expression of the process is as follows:

[0060]

[0061] wherein, represents the original position of the dung beetle population, represents the solution obtained by using the Tent mapping; represents the control parameter of the Tent mapping; represents the reverse solution of ; and represent the upper and lower limits of the position of the dung beetle population respectively, represents the random reverse solution, and rand represents a random number between 0 and 1.

[0062] The expression of the candidate dung beetle population set is as follows:

[0063] wherein, C represents the candidate dung beetle population set.

[0064] Exemplarily, is the control parameter of the Tent mapping.

[0065] The Tent map possesses more complex nonlinear properties, generating chaotic sequences with more uniform distribution, lower correlation, and better ergodicity. Furthermore, an improved stochastic back-learning strategy is introduced when initializing the population using the Tent map to enhance the quality of population-based solutions. By balancing global exploration (Tent map) and local exploitation (back-learning) from the initial stage, the algorithm helps escape local optima and improves its global search capability.

[0066] like Figure 2 The top view of the drone trajectory planning results shown is as follows: Figure 3 The standard view shown represents the path generated by the UAV trajectory planning method based on the multi-strategy improved dung beetle optimization algorithm, on a two-dimensional plane. Figure 2 It exhibits smooth and highly obstacle-avoiding curved characteristics, avoiding obstacles in multiple areas. In three-dimensional space ( Figure 3 The trajectory precisely avoids the top areas of three-dimensional obstacles such as hills, always maintaining a safe flight altitude. It achieves global optimality and smoothness of the path, ensuring a more optimized flight path length and gentler turns for the drone. Through the deep integration of 3D obstacle modeling and performance constraints, reliable obstacle avoidance in complex terrain is achieved, validating its significant advantages in balancing path optimization and safe obstacle avoidance.

[0067] Example 2 In this embodiment, the multi-strategy improved dung beetle optimization algorithm mentioned in Embodiment 1 is described in detail. The process of updating the position of individual dung beetles using the multi-strategy method is as follows: the position of the rolling dung beetle is updated using the rolling dung beetle optimization algorithm; the position of the reproductive dung beetle is updated using the reproductive dung beetle optimization algorithm; the position of the foraging dung beetle is pre-updated using the foraging dung beetle optimization algorithm; the updated position of the foraging dung beetle is calculated and updated using the subtractive averaging strategy; the position of the thief dung beetle is pre-updated using the thief dung beetle optimization algorithm; and the position of the thief dung beetle is updated using the adaptive t-distribution perturbation strategy and Cauchy mutation.

[0068] An optimization algorithm for the rolling dung beetle is used to update the individual positions of the rolling dung beetle; the rolling behavior of the rolling dung beetle includes obstacle-prone and obstacle-free modes. In obstacle mode, the expression for updating the position of an individual dung beetle is:

[0069] In accessibility mode, the expression for updating the position of an individual dung beetle is:

[0070] In the formula, t Indicates the current iteration number. This represents the position of individual dung beetle n at iteration t. represents a natural number, k represents a deflection coefficient, b represents a constant between 0 and 1, represents an absolute distance between the current roller ball individual and the global worst roller ball individual, represents a global worst position, represents a position difference of the n-th roller ball individual at the t-th iteration and at the (t-1)-th iteration; exemplary, is a natural number set as -1 or 1, k∈(0, 0.2] is a deflection coefficient, b∈(0, 1) is a constant between 0 and 1. When θ is 0, or π, the roller ball individual position will not be updated.

[0071] The roller ball individual position is updated by using the breeding roller ball optimization algorithm; the breeding roller ball area includes an oviposition area and a breeding area; In the oviposition area, the expression for updating the breeding roller ball individual position is:

[0072] In the breeding area, the expression for updating the breeding roller ball individual position is:

[0073] In the formula, represents a local optimal roller ball individual, and represents the upper and lower bounds of the simulated oviposition area, represents a maximum iteration number, and represents the lower and upper limits of the search space; represents a position of the brood ball n at the t-th iteration, and represents a random vector, D represents a dimension of an optimization problem, and R represents a boundary convergence factor; In the embodiment, the foraging roller ball individual position is updated in advance by using the foraging roller ball optimization algorithm; the expression for the foraging roller ball individual in the optimal foraging area is:

[0074] When the foraging roller ball individual forages, the expression for updating the foraging roller ball individual position is:

[0075] In the formula, represents a global optimal solution, and represents the lower and upper limits of the optimal foraging area; represents a position of the roller ball individual n at the t-th iteration, a random number obeying normal distribution, a random vector; exemplary, a random vector in the range of (0, 1).

[0076] In each iteration, the individual will be updated based on the subtraction formula and the average of all other individuals, and the position will be updated as the offset.

[0077] The update position of the foraging scarab individual is calculated by using the subtraction average strategy, and the expression is:

[0078] In the formula, represents the update position of the foraging scarab individual, represents a vector of dimension m, represents the position of the foraging scarab individual j; represents the i-th vector of dimension m, whose components are taken from the interval [0, 1] and obey normal distribution.

[0079] According to the update position of the foraging scarab individual, the position of the foraging scarab individual is updated, and the expression is:

[0080] In the formula, represents the position of the foraging scarab individual after updating, represents the weight coefficient, represents the weight coefficient, represents the upper limit of the best foraging area, represents a vector of dimension m, represents the lower limit of the best foraging area, represents the position of the foraging scarab individual j at the t-th iteration, represents the subtraction average operation rule, N represents the total number of individuals in the scarab population; In this embodiment, the subtraction average strategy proposes a new way of calculating the difference value "subtraction" ". The definition expression of "subtraction" is:

[0081] In the formula, A and B represent two population individuals of dimension m, and represent the fitness value; represents an m-dimensional vector, and the components are randomly generated from the set {1, 2}.​ denotes the Hadamard product between two vectors.

[0082] The population position of the forager is updated by a subtraction average strategy, and the position of the forager is updated based on the arithmetic average position instead of the optimal position, which can effectively explore and utilize the best foraging area, thereby avoiding premature convergence.

[0083] The position of the thief is updated in advance using the thief optimization algorithm, and the expression is:

[0084] In the formula, denotes the optimal food source, nearby indicates the best place for the thief to compete for food, g denotes a random vector subject to normal distribution, and s denotes a constant; For example, the size of g is set to 1 x D.

[0085] The position of the thief is disturbed by introducing an adaptive t-distribution, which is used to increase the population diversity of the thief, and the expression of the adaptive t-distribution disturbance strategy is:

[0086] In the formula, m denotes the degree of freedom parameter, denotes the gamma function; The target position of the thief is updated by introducing Cauchy variation, which improves the global optimization performance of the algorithm, and the expression of the Cauchy variation is:

[0087] In the formula, denotes the Cauchy variation.

[0088] In each iteration process, for each individual in the population, a new candidate solution is generated based on the current global optimal solution. The generation of the new solution is realized by the following way: starting from the global optimal solution, the disturbance is carried out along the random direction, and the disturbance direction is determined by the combination of adaptive t-distribution and Cauchy distribution.

[0089] The position of the thief individual is updated by using the adaptive t-distribution disturbance strategy and Cauchy variation, and the expression is:

[0090] In the formula, denotes the position of the thief individual after adaptive t-distribution disturbance, denotes the optimal position of the current thief individual; denotes the random number generated by the adaptive t-distribution; denotes the random number generated by the Cauchy distribution, and a contribution coefficient representing the adaptive t-distribution and the Cauchy distribution; exemplarily, and range between (0, 1).

[0091] Adopting the fusion adaptive t-distribution disturbance strategy and the Cauchy variation to update the population position of the thief scarab, the diversity of the scarab population is effectively maintained, the premature convergence is avoided, and the search performance of the thief scarab in the scarab optimization algorithm is significantly improved, so that the algorithm performs better when solving complex optimization problems.

[0092] Adopting the subtraction average strategy, the fusion adaptive t-distribution disturbance strategy and the Cauchy variation to update the population position of the scarab, the number of invalid evaluations is reduced, the algorithm can locate the global optimal region faster and more accurately, thereby improving the search efficiency, maintaining the diversity of the scarab population and avoiding premature convergence.

[0093] Embodiment 3 In this embodiment, steps S4 and S5 of the unmanned aerial vehicle trajectory planning method based on the multi-strategy improved scarab optimization algorithm mentioned in embodiment 1 are described in detail.

[0094] The expression of the target cost function constructed by the length cost constraint, the height cost constraint, the yaw angle constraint and the pitch angle constraint of the unmanned aerial vehicle is:

[0095]

[0096] In the formula, F F represents the target cost function, , , , represents the evaluation weight. When finding the best position of the scarab individual, the target cost function F, that is, the flight task cost, is needed to calculate the fitness value of the scarab individual. There are N scarabs in the population, and the fitness value of each scarab is calculated. The smallest fitness value is the best position of the corresponding scarab individual Using the obstacle model in the three-dimensional environment of the unmanned aerial vehicle and the best position of the scarab individual, the global path curve of the unmanned aerial vehicle is obtained by using spline interpolation. The process is as follows: the intersection points of the spline interpolation are taken as the path nodes, and all the path nodes of a path form a set; the three-dimensional coordinates of the path nodes are split into three sets according to the dimensions; the three sets are merged to obtain the coordinates of the best position of the scarab individual; based on the coordinates of the best position of the scarab individual, the starting point and the target point are set to obtain the global path curve of the unmanned aerial vehicle.

[0097] Exemplarily, the number of spline interpolation is set to 3. In the embodiment, the obstacle model in the three-dimensional environment of the UAV is used to obtain , describes the UAV path curve. Cubic spline interpolation is used for the intersection points of each segment as the path node, and all path nodes of a path are used as the individual optimal position of the Scarabaeidae. There are m path nodes on a path, and the expression of the path node coordinate set is:

[0098] In the formula, represents the position coordinates of the Scarabaeidae, and R represents the path node coordinate set.

[0099] The three-dimensional coordinates of the path node are divided into three sets according to the dimensions, and the expressions are respectively:

[0100]

[0101] .

[0102] The path node three-dimensional coordinate sets , and are combined as the optimal position coordinates of the Scarabaeidae, and the expression is:

[0103] In the formula, x represents the optimal position coordinates.

[0104] The dimension of the optimal position of the Scarabaeidae is 3m, and the dimension of the optimal position of the Scarabaeidae population is three times the number of path nodes. At the same time, in combination with the set starting point S to the target point G , the coordinates of n interpolation points can be obtained by the method of cubic spline interpolation, and the connecting line of the interpolation point coordinates is the planned UAV global path curve.

[0105] The embodiments are only examples for clearly illustrating the present application, and are not limitations on the embodiments of the present application. Based on the above description, other different forms of changes or variations can be made by those of ordinary skill in the art. Here, all embodiments are not required to be exhausted. Any modification, equivalent replacement and improvement made within the spirit and principle of the present application shall be included in the protection scope of the claims of the present application.

Claims

1. A method for UAV trajectory planning based on a multi-strategy improved dung beetle optimization algorithm, characterized in that, Includes the following steps: S1: Establish an obstacle model in the 3D environment of the UAV and design UAV performance constraints; S2: Construct a dung beetle population containing several dung beetle individuals; use a fusion of Tent mapping and random back-learning strategy to initialize the dung beetle population and generate the initial dung beetle individual locations; S3: Employ a multi-strategy improvement algorithm for dung beetle optimization to update the individual dung beetle positions; S4: Based on the performance constraints of the drone designed in step S1, construct the objective cost function, calculate the fitness value of each dung beetle individual, and update the optimal position of the dung beetle individual; S5: Using the obstacle model in the UAV's 3D environment and the optimal position of the dung beetle, spline interpolation is used to obtain the global path curve of the UAV; S6: Repeat steps S2-S5 until the preset maximum number of iterations is reached, and output the final optimal trajectory of the UAV.

2. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 1, characterized in that, The expression for the obstacle model in the UAV's 3D environment is: In the formula, The function representing the height of the hill. Represents the coordinates of a point in two-dimensional space. Indicates the center coordinates of the hill. h The parameter representing the height of the hill. and This represents the steepness parameter of the hill.

3. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 1, characterized in that, The performance constraints of the UAV include: length cost constraint, altitude cost constraint, yaw angle constraint, and pitch angle constraint; The expression for the length cost constraint is as follows: In the formula, Represented as the i-th path node of the UAV With the (i+1)th path node The distance between them This represents the position values ​​of the i-th path node of the UAV on the x-axis, y-axis, and z-axis. This represents the position values ​​of the (i+1)th path node of the drone on the x, y, and z axes. Indicates length cost constraint; The expression for the high cost constraint is: In the formula, Indicates the minimum flight altitude of the drone. Indicates the maximum flight altitude of the drone. This indicates the flight altitude of the drone at waypoint i. Indicates the cost of drone path altitude; The expression for the yaw angle constraint is: In the formula, This indicates the yaw angle constraint. Indicates the yaw angle. and This indicates the minimum and maximum yaw angles of the drone during flight. The vector representing the i-th path segment; The expression for the pitch angle constraint is: In the formula, This indicates the yaw angle constraint. Indicates pitch angle, This indicates the maximum climb angle of the drone.

4. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 1, characterized in that, Dung beetle populations include: rolling dung beetle, reproductive dung beetle, foraging dung beetle, and thieving dung beetle.

5. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 1, characterized in that, The process of initializing the dung beetle population and generating initial dung beetle individual locations using a fusion of Tent mapping and stochastic backpropagation strategies is as follows: The original solution for the location of individual dung beetles is generated by fusion Tent mapping; and the reverse solution for the location of individual dung beetles with random perturbation is generated based on the original location of individual dung beetles using a stochastic reverse learning strategy. Based on the original solution and the reverse solution of the dung beetle individual positions, a candidate dung beetle population set C is constructed. The fitness value F of each dung beetle individual in the candidate dung beetle population set C is calculated iteratively, and the n dung beetle individuals with the best fitness value F are selected to form the initial dung beetle individual positions. The expression for this process is: In the formula, Indicates the original location of the dung beetle population. express The solution obtained using Tent mapping; Indicates the control parameters for Tent mapping; express The inverse solution; and These represent the upper and lower limits of the dung beetle population location, respectively. This represents the inverse solution of a random number, and rand represents a random number between 0 and 1.

6. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 4, characterized in that, The process of updating the individual positions of dung beetles using a multi-strategy approach is as follows: updating the individual positions of dung beetles using the rolling dung beetle optimization algorithm; updating the individual positions of breeding dung beetles using the reproductive dung beetle optimization algorithm; and updating the individual positions of foraging dung beetles in advance using the foraging dung beetle optimization algorithm. A subtractive averaging strategy is used to calculate and update the updated positions of foraging dung beetles. The dung beetle thief optimization algorithm is used to pre-update the positions of dung beetle thief individuals. An adaptive t-distribution perturbation strategy and Cauchy mutation are used to update the positions of dung beetle thief individuals.

7. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 6, characterized in that, An optimization algorithm for the rolling dung beetle is used to update the individual positions of the rolling dung beetle; the rolling behavior of the rolling dung beetle includes obstacle-prone and obstacle-free modes. In obstacle mode, the expression for updating the position of an individual dung beetle is: In accessibility mode, the expression for updating the position of an individual dung beetle is: In the formula, t Indicates the current iteration number. This represents the position of individual dung beetle n at iteration t. Let k represent a natural number, b represent the deflection coefficient, and b represent a constant between them. This represents the absolute distance between the current dung beetle individual and the worst dung beetle individual globally. Indicates the worst position globally. This represents the difference between the position of the nth dung beetle individual at the t-th iteration and its position at the (t-1)th iteration; The location of individual dung beetles is updated using a breeding dung beetle optimization algorithm; The area where the breeding dung beetles are located includes: an egg-laying area and a breeding area; In the egg-laying area, the expression for updating the location of reproductive dung beetles is: In the breeding zone, the expression for updating the location of breeding dung beetles is: In the formula, This represents the locally optimal dung beetle individual. and use This indicates the upper and lower boundaries of the simulated spawning area. Indicates the maximum number of iterations. and Indicates the lower and upper limits of the search space; This represents the position of the brood ball n in the t-th iteration. and Let represent a random vector, D represent the dimension of the optimization problem, and R represent the boundary convergence factor. Using a dung beetle optimization algorithm, the positions of individual dung beetles are pre-updated; the expression for an individual dung beetle's optimal foraging area is: The expression for updating the position of a foraging dung beetle when it is foraging is: In the formula, This represents the globally optimal solution. and Indicates the lower and upper limits of the optimal foraging area; This represents the position of individual dung beetle n at iteration t. Represents a random number that follows a normal distribution. Represents a random vector; The updated position of a foraging dung beetle is calculated using a subtractive averaging strategy, expressed as: In the formula, This indicates the updated location of a foraging dung beetle individual. Let m represent a vector of dimension m. Indicates the position of individual dung beetle j; Update the position of the foraging dung beetle based on its updated position. The expression is: In the formula, This indicates the location of the foraging dung beetle after the update. Indicates the weighting coefficient. Indicates the weighting coefficient. This indicates the upper limit of the optimal foraging area. Let m represent a vector of dimension m. This indicates the lower limit of the optimal foraging area. This represents the position of dung beetle individual j at iteration t. This indicates the rules for averaging by subtraction. N This indicates the total number of individuals in a dung beetle population. An optimization algorithm for dung beetles is used to pre-update the positions of individual dung beetles; its expression is: In the formula, Let g represent the optimal food source, g represent a random vector that follows a normal distribution, and s represent a constant. Using an adaptive t-distribution perturbation strategy and Cauchy mutation, the position of individual dung beetles is updated, expressed as: In the formula, This represents the position of an individual dung beetle after adaptive t-distribution perturbation. This indicates the optimal position of the current dung beetle individual; This represents a random number generated by an adaptive t-distribution; This represents a random number generated by the Cauchy distribution. and This represents the contribution coefficients of the adaptive t-distribution and the Cauchy distribution.

8. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 7, characterized in that, The expression for the adaptive t-distribution perturbation strategy is: In the formula, m represents the degree of freedom parameter. Represents the gamma function; The expression for the Cauchy variation is: In the formula, This indicates a Cauchy mutation.

9. The UAV trajectory planning method based on a multi-strategy improved dung beetle optimization algorithm according to claim 3, characterized in that, The expression for the objective cost function is: In the formula, F Represents the objective cost function. , , , Indicates the evaluation weight.

10. A method for UAV trajectory planning based on a multi-strategy improved dung beetle optimization algorithm according to claim 3, characterized in that, The process of obtaining the UAV's global path curve using spline interpolation based on the obstacle model in the UAV's 3D environment and the optimal position of the dung beetle is as follows: the intersection points of the spline interpolation segments are taken as path nodes, and all path nodes of a path form a set; the 3D coordinates of the path nodes are split into three sets according to the dimensions; the three sets are merged to obtain the optimal position coordinates of the dung beetle; based on the optimal position coordinates of the dung beetle, combined with the set starting point and target point, the global path curve of the UAV is obtained.

Images