Unmanned aerial vehicle cluster target strike route planning method based on optimization algorithm

Abstract

The invention discloses an unmanned aerial vehicle cluster target strike route planning method based on an optimization algorithm. The method comprises the following steps: (1) obtaining a three-dimensional complex urban environment model by utilizing a dimensional elevation map modeling method and MATLAB processing according to complex urban environment characteristics; (2) establishing unmanned aerial vehicle constraint conditions oriented to the complex urban environment model; (3) establishing an unmanned aerial vehicle track cost function oriented to the complex urban environment model, and obtaining a track evaluation result; and (4) obtaining the optimal route of the flight path planning of the unmanned aerial vehicle cluster in the complex urban environment by adopting an optimization algorithm. According to the method, a reasonable target strike trajectory is planned for the unmanned aerial vehicle by adopting the optimization algorithm based on the complex urban environment model, the unmanned aerial vehicle constraint condition and the flight path evaluation result in the complex urban environment, so that the flight path planning of the unmanned aerial vehicle has higher flight path planning efficiency and stability, the consumed flight path cost is minimum, and convergence is faster.

Description

Technical Field

[0001] The present invention belongs to the technical field of trajectory planning, and in particular relates to a method for planning a target strike trajectory of a cluster of unmanned aerial vehicles based on an optimization algorithm. Background Art

[0002] One of the key technologies for UAV urban combat is UAV trajectory planning in urban environments. Currently, there is little research on UAV trajectory planning in urban environments. et al. adopted the Coverage Path Planning (CPP) algorithm for path planning, but it requires a comprehensive exploration of the urban environment, which is time-consuming and lacks real-time performance. In contrast, the Snake Optimizer (SO) algorithm is more efficient in processing real-time path generation. Jianfa Wu et al. proposed a path planning framework for Solar Unmanned Aerial Vehicle (SUAV), which solved the problem of accurate solar energy modeling, but this method has limitations in planning in dim environments, while the Sine cosine algorithm (SCA) performs stably under a variety of lighting conditions. Xinting Hu et al. proposed a path planning method based on risk cost. Although it takes into account a variety of risks, it relies on a preset environmental model and lacks real-time performance. Therefore, it is a hot topic to study how to avoid relying on preset environmental models in the path planning of UAVs in urban environments and achieve a method that is highly adaptable, time-saving, and highly real-time. Summary of the invention

[0003] In view of the above-mentioned deficiencies in the prior art, the present invention invents a method for planning the trajectory of a swarm of unmanned aerial vehicles (UAVs) against target attacks based on an optimization algorithm.

[0004] In order to achieve the above object, the present invention adopts the following technical solution:

[0005] A method for UAV cluster target attack trajectory planning based on optimization algorithm comprises the following steps:

[0006] (1) According to the characteristics of complex urban environment, a three-dimensional complex urban environment model is obtained by using the three-dimensional elevation map modeling method and MATLAB processing;

[0007] (2) Establishing UAV constraint conditions for a complex urban environment model, wherein the UAV constraint conditions include safe flight altitude area constraint, maximum range constraint, and waypoint obstacle avoidance area constraint;

[0008] (3) Establish a UAV trajectory cost function for a complex urban environment model, evaluate the trajectory of the UAV cluster, and obtain the trajectory evaluation results;

[0009] (4) An optimization algorithm is used to combine the complex urban environment model, UAV constraints and trajectory evaluation results to obtain the optimal route for trajectory planning of UAV clusters in complex urban environments.

[0010] Furthermore, the complex urban environment model includes urban building danger area modeling and threat area modeling, and the urban building danger area modeling and threat area modeling are constructed by utilizing the three-dimensional elevation map modeling method to truly restore the information of urban building topography, threat areas and other enemy defense measures.

[0011] Furthermore, the modeling of the dangerous area of ​​urban buildings is to set the urban buildings as cylinders, and the drone is in a dangerous state when it is in this area. The modeling expression formula is:

[0012]

[0013] Among them, the parameters of the nth cylindrical city building are expressed as (a 1n , b 1n ,h 1n , R 1n ), (a 1n , b 1n ) represents the coordinates of the building center area, h 1n Represents the building height, R 1n represents the radius corresponding to the building, 1 represents the building, n is a non-zero natural number, and (x, y, z) is the coordinate of the drone track point.

[0014] Furthermore, the threat area modeling is to set the threat area as a hemisphere, and the modeling expression formula is:

[0015]

[0016] The mth ground threat area is represented by (a 2m , b 2m , R 2m ), (a 2m , b 2m ) represents the coordinates of the threat center area, 2 represents the threat point, m is a non-zero natural number, R 2m Indicates the radius corresponding to the threat point.

[0017] Furthermore, the safe flight altitude area is constrained as follows:

[0018] H min ≤H≤H max

[0019] Among them, H represents the current flight altitude of the drone, H min is the minimum safe flight altitude, H maxThe maximum safe flight altitude;

[0020] The maximum range constraint is:

[0021]

[0022] Among them, l max Indicates the maximum value of the flight range, l e Represents each flight segment, and the total flight line number of the drone is {l 1 , l 2 , l 3 …l e}A total of e track lines;

[0023] The waypoint obstacle avoidance area constraints are:

[0024] z n >A(x, y, z) or z m >B(x,y)

[0025] Among them, z n represents the elevation of the UAV track point n, z m Indicates the elevation of the UAV track point m.

[0026] Furthermore, the trajectory of the drone cluster is evaluated as follows:

[0027] (1) Setting performance indicators for the trajectory planning of UAV swarm target attack: range cost, altitude change cost, and turning angle cost;

[0028] (2) The track cost function is obtained through the range cost function, altitude change cost function, and turning angle cost function:

[0029] J cost =D cost +H cost +Q cost

[0030] Among them, J cost is the total track cost function, D cost is the flight cost function, H cost is the height variation cost function, Q cost is the turning angle cost function;

[0031] The range cost function is the actual flight distance of the drones in the drone cluster from track point k to track point k+1, and its expression is:

[0032]

[0033] The height change cost function is:

[0034]

[0035] Where C is the number of track points, is the average height of the track points, and Z is the flight altitude of the drone;

[0036] The turning angle cost function is:

[0037]

[0038] By solving the track cost function, the track composed of line segments is obtained, and then the cubic B-spline method is used to smooth it, and finally the evaluation result of the actual flyable track of the UAV is obtained.

[0039] Further, the specific steps of the optimization algorithm are as follows:

[0040] S1: Automatically obtain the starting point coordinates, target point coordinates, complex urban environment model, UAV constraint conditions, trajectory evaluation results, maximum number of iterations T and UAV population number g in the UAV trajectory planning control system;

[0041] S2: Calculate the energy factor E. The formula for calculating E is:

[0042]

[0043] Where t represents the current number of iterations, T represents the maximum number of iterations, and r is a random number in (0, 1);

[0044] If E>1, each drone represents a candidate solution, and the best solution in each step of the drone is regarded as the coordinates of the target point. Each drone interacts with other drones through position information and can interact with any drone in the fleet. The interaction between drones is marked by stirring. Then the drone position is updated according to the navigation interaction behavior. The calculation formula is:

[0045]

[0046] Among them, during the navigation phase, V i (t+1) represents the position vector of UAV i at the t+1th iteration, V j (i) represents the position vector of drone j at the tth iteration, j is the position of a drone randomly selected from the current drone group and j≠i, F represents a random matrix that can randomly generate 0 and 1, represents the average position vector of the UAV, p 1 Represents a random number within (0, 1);

[0047] If 2 / 3<E<1, the drone explores a new area:

[0048] V i (t+1)=-r×Ui (t+1)+r×H r (t+1)-L×(H r (t+1)-V i (t))

[0049] Among them, in the exploration stage, H r (t+1) represents the random position of the drone in the exploration area, r is a random number in (0, 1), and L is the Levy flight function used to expand the search space;

[0050] If 1 / 3<E<2 / 3, the UAV is corrected and locally optimized to approach the target path. The expression formula is:

[0051] U i (t+1)=W(t)+α×|W(t)-x prey (t)|

[0052]

[0053] Among them, in the correction stage, U i (t+1) represents the corrected position vector of UAV i in the t+1 iteration, W(t) represents the projected position vector of the current UAV in the normalized search space, x prey (t) represents the current optimal solution position of the target UAV, α represents the initial scale of the resting zone, and controls the strength of local optimization;

[0054] According to the modified behavior, the drone updates its navigation position according to the formula:

[0055] V i (t+1)=U i (t+1)+n 1 ×(U i (t+1)round(rand)×V i (t))

[0056] Among them, n 1 Indicates a random number in (0, 1), and round(rand) indicates generating a random number of 0 or 1.

[0057] Machine function;

[0058]

[0059] In the attack phase, H prey (t+1) represents the position of the target UAV in the t+1th iteration, β is the reduction coefficient of the control lock area, and the scale β makes the range of the strike area smaller over time.

[0060] If 0<E<1 / 3, update the position according to the locking behavior, and the calculation formula is:

[0061] V i (t+1)=H prey (t+1)+η×(H prey (t+1)-round(rand)×V i (t))

[0062] Where η represents the curling factor;

[0063] S3: The position of the drone is based on the fitness fit (V i (t)) to decide whether to update:

[0064]

[0065] Among them, fit(Vi(t)) represents the fitness of the current UAV’s track;

[0066] S4: After each iteration, update the current optimal solution x prey (t) Obtain the optimal path in the UAV trajectory.

[0067] The beneficial effects of the present invention are as follows: the present invention plans a reasonable target strike trajectory for the UAV based on the complex urban environment model, the UAV constraint conditions and the track evaluation results in a complex urban environment by using an optimization algorithm, so that the UAV track planning has higher track planning efficiency and stability, the consumed track cost is minimized and the convergence is faster, so as to achieve the optimal UAV target strike trajectory and a better convergence trend of the consumed track cost, the highest task completion and better performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 It is a flow chart of a method for planning the trajectory of a drone cluster target attack based on an electric eel foraging optimization algorithm in a complex urban environment according to the present invention;

[0069] Figure 2 A schematic diagram of urban buildings provided by the present invention;

[0070] Figure 3 A schematic diagram of the threat area provided by the present invention;

[0071] Figure 4 A comparison chart of flight trajectories of EEFO, HHO, MFO, SCA, SO, and WOA algorithms provided by the present invention for complex urban environments;

[0072] Figure 5 A line comparison chart of the optimal fitness value of the algorithm provided by the present invention;

[0073] Figure 6A bar graph showing the difference in the optimal fitness values ​​of the algorithm provided by the present invention;

[0074] Figure 7 This is a thermal comparison diagram of the optimal fitness value of the algorithm provided by the present invention. DETAILED DESCRIPTION

[0075] The present invention is described in detail below with reference to the accompanying drawings:

[0076] A method for UAV cluster target attack trajectory planning based on optimization algorithm includes the following steps:

[0077] (1) According to the characteristics of complex urban environment, a three-dimensional complex urban environment model is obtained by using 3D elevation map modeling and MATLAB processing; the complex urban environment model includes urban building danger area modeling and threat area modeling, which are constructed by using 3D elevation map modeling to truly restore the information of urban building topography, threat area and other enemy defense measures. Urban building danger area modeling not only retains the main limiting factors of buildings on drone flight, but also significantly reduces the computational complexity, so that the drone path planning algorithm can process more efficiently and find a safe detour path. Threat area modeling can simulate the full coverage of ground threats and their action areas with high fidelity.

[0078] (2) Establish UAV constraint conditions for a complex urban environment model, wherein the UAV constraint conditions include safe flight altitude area constraint, maximum range constraint, and waypoint obstacle avoidance area constraint.

[0079] (3) Establish a UAV track cost function for a complex urban environment model, evaluate the track of the UAV cluster, and obtain the track evaluation results.

[0080] (4) An optimization algorithm is used to combine the complex urban environment model, UAV constraints and trajectory evaluation results to obtain the optimal route for trajectory planning of UAV clusters in complex urban environments.

[0081] like Figure 2 As shown in the figure, the modeling of the dangerous area of ​​urban buildings is to set the urban buildings as cylinders. The drone is in a dangerous state when it is in this area. The modeling expression formula is:

[0082]

[0083] Among them, the parameters of the nth cylindrical city building are expressed as (a 1n , b 1n ,h 1n , R 1n ), (a 1n , b 1n) represents the coordinates of the building center area, h 1n Represents the building height, R 1n Represents the radius corresponding to the building, 1 represents the building, n is a non-zero natural number, and (x, y, z) is the coordinates of the drone's track point. By modeling urban buildings, drones can effectively avoid or discover new building danger zones through model calculations, and ultimately achieve optimization of the navigation path.

[0084] like Figure 3 As shown, the threat area modeling is to set the threat area as a hemisphere, and the modeling expression formula is:

[0085]

[0086] The mth ground threat area is represented by (a 2m , b 2m ,R 2m ), (a 2m , b 2m ) represents the coordinates of the threat center area, 2 represents the threat point, m is a non-zero natural number, R 2m Indicates the radius corresponding to the threat point. The threat area is composed of ground-deployed radar systems, electromagnetic interference devices, missile vehicles, etc. The hemispherical model, as an analysis tool, can simulate the full coverage of ground threats and their action areas with high fidelity.

[0087] When performing precision strike missions in urban environments, although a high flight altitude can expand the field of view, it increases the risk of being detected by enemy radars; a low altitude may make it difficult for the drone to deal with sudden obstacles, increasing the risk of collision. Therefore, the safe flight altitude area constraints are:

[0088] H min ≤H≤H max

[0089] Among them, H represents the current flight altitude of the drone, H min is the minimum safe flight altitude, H max is the maximum safe flight altitude. So the maximum safe flight altitude H max The lowest altitude detected by enemy radar, the minimum safe flight altitude H min Should be lower than h 1n Or the height that does not belong to any point in B(x, y).

[0090] In addition to basic energy consumption, the complexity of maneuvering, the unpredictability of wind resistance, and variable environmental factors such as temperature and air pressure will further increase energy consumption, thereby significantly reducing the endurance of the drone. In order to optimize the combat effectiveness of the drone and extend its mission execution time, it is particularly important to reasonably set the range constraint of the drone. The maximum range constraint is:

[0091]

[0092] Among them, l max Indicates the maximum value of the flight range, l e Represents each flight segment, and the total flight line number of the drone is {l 1 , l 2 , l 3 …l e}It consists of e track lines in total.

[0093] To ensure that the drone can maintain a safe distance when flying through the city and avoid any form of collision risk, the waypoint obstacle avoidance area constraints are:

[0094] z n >A( x， y, z) or z m >B( x， y)

[0095] Among them, z n represents the elevation of the UAV track point n, z m Indicates the elevation of the drone track point m. In this way, the drone can accurately avoid tall buildings and potential threat areas in the city.

[0096] In the trajectory planning of UAVs, the range cost is a key decision factor. The turning angle directly affects the flight time, energy efficiency and mission execution effect. Therefore, the optimization strategy usually focuses on minimizing the flight distance between the UAV track points. Therefore, the trajectory evaluation of the UAV cluster is as follows:

[0097] (1) Setting performance indicators for the trajectory planning of UAV swarm target attack: range cost, altitude change cost, and turning angle cost;

[0098] (2) The track cost function is obtained through the range cost function, altitude change cost function, and turning angle cost function:

[0099] J cost =D cost +H cost +Q cost

[0100] Among them, J cost is the total track cost function, D cost is the flight cost function, H cost is the height variation cost function, Q cost is the turning angle cost function. The total track cost function J cost The smaller the value, the better the trajectory planning optimization.

[0101] The range cost function is the actual flight distance of the drones in the drone cluster from track point k to track point k+1, and its expression is as follows through the Euclidean distance:

[0102]

[0103] By calculating the standard deviation of altitude during the flight, the flight stability is evaluated and a basis is provided for optimizing the strategy and reducing unnecessary altitude adjustments. The altitude change cost function is:

[0104]

[0105] Where C is the number of track points, is the average height of the track points, and Z is the flight altitude of the drone;

[0106] The turning angle cost function is used as an evaluation indicator to measure the turning frequency and smoothness of the path, so the turning angle cost function is:

[0107]

[0108] By solving the track cost function, the track composed of line segments is obtained, and then the cubic B-spline method is used to smooth it, and finally the evaluation result of the actual flyable track of the UAV is obtained.

[0109] The specific steps of the optimization algorithm are as follows:

[0110] S1: Automatically obtain the starting point coordinates, target point coordinates, complex urban environment model, UAV constraint conditions, trajectory evaluation results, maximum number of iterations T and UAV population number g in the UAV trajectory planning control system;

[0111] S2: Calculate the energy factor E. The formula for calculating E is:

[0112]

[0113] Among them, t represents the current number of iterations, T represents the maximum number of iterations, and r is a random number in (0, 1).

[0114] The introduction of energy factor is used to balance global and local search. In the early stage, the energy factor is higher, which promotes global search. As the iteration proceeds, the energy factor gradually decreases, focusing more on local search and improving the accuracy of the solution.

[0115] If E>1, then the drone position is updated according to the navigation interaction behavior, and the calculation formula is:

[0116]

[0117] in,

[0118] Among them, during the navigation phase, V i (t+1) represents the position vector of UAV i at the t+1 iteration, and the next position is determined by the interaction between UAVs; V j (t) represents the position vector of drone j at the tth iteration, where j is the position of a drone randomly selected from the current drone group - and j≠i; F represents a random matrix that can randomly generate 0 and 1; Represents the average position vector of the UAV; V r (t) represents the position of a UAV randomly selected from the current fleet; p 1 Represents a random number in (0, 1), which is used to decide whether the next updated position is based on the average position or the position of another randomly selected drone. Low and Up represent the lower and upper limits of the solution space, respectively, and are used to define the boundaries of the drone search.

[0119] If 2 / 3<E<1, the drone explores a new area:

[0120] V i (t+1)=-r×U i (t+1)+r×H r (t+1)-L×(H r (t+1)-V i (t))

[0121] Among them, in the exploration stage, H r (t+1) represents the random position of the UAV in the exploration area, r is a random number in (0, 1), and L is the Levy flight function, which has a heavy-tailed distribution and can generate long-distance jumps to avoid the search from falling into the local optimum, and is used to expand the search space.

[0122] Among them, H r (t+1) is based on x in the target area prey (t) and the corrected scale coefficient are calculated. The specific calculation formula is:

[0123]

[0124] β=β0xsin(2 π r)

[0125]

[0126] β represents the scale factor of the hunting area, which controls the offset of the target position. β decreases gradually over time, making the search range gradually converge. 0 is the initial scale factor.

[0127] If 1 / 3<E<2 / 3, the UAV is corrected and locally optimized to approach the target path. The expression formula is:

[0128] Ui(t+1)=W(t)+α×|W(t)-x prey (t)|

[0129] α=α 0 ×sin(2πr)

[0130]

[0131] Among them, in the correction stage, U i (t+1) represents the corrected position vector of UAV i in the t+1 iteration, W(t) represents the projected position vector of the current UAV in the normalized search space, x prey (t) represents the current optimal solution position of the target UAV, α represents the initial scale of the resting zone, and controls the strength of local optimization; α 0 Beginning scale.

[0132] According to the modified behavior, the drone updates its navigation position according to the formula:

[0133] V i (t+1)=U i (t+1)+n 1 ×(U i (t+1)round(rand)×V i (t))

[0134] Among them, n 1 represents a random number in (0, 1). This random perturbation term makes the position of the drone fluctuate slightly during the correction phase to avoid excessive concentration. round(rand) represents a random function that generates 0 or 1, which is used to randomly decide whether to introduce the current drone position V i (t), if round(rand) = 1, it will include the information of the current UAV position, if round(rand) = 0, it will include the U i (t+1) information.

[0135] If 0<E<1 / 3, update the position according to the locking behavior, and the calculation formula is:

[0136] V i (t+1)=H prey (t+1)+η×(H prey (t+1)-round(rand)×V i (t))

[0137] Among them, η represents the curl factor, which is used to control the random perturbation intensity when the UAV position is updated, and it changes with the advancement of time t;

[0138]

[0139] r is a random number (0, 1), which determines the specific value of the curl factor and ensures the randomness and curl characteristics of the drone path.

[0140] S3: The position of the drone is based on the fitness fit (V i (t)) to decide whether to update:

[0141]

[0142] Among them, fit(V i (t)) represents the current UAV’s track fitness;

[0143] S4: After each iteration, update the current optimal solution x prey (t) Obtain the optimal path in the UAV trajectory.

[0144] like Figure 4 As shown, this embodiment uses five algorithms for comparison, namely EEFO (the calculation method of the present invention), SCA (sine and cosine optimization algorithm), S0 (snake optimization algorithm), WOA (whale optimization algorithm), MFO (moth flame optimization algorithm) and HHO (Harris Hawk optimization algorithm). In the dense urban environment with enemy defense scenarios, facing the challenge of the enemy's tight defense, the target strike mission of the drone is particularly arduous. The EEFO algorithm shows extraordinary navigation and avoidance capabilities. The algorithm can not only accurately guide the drone to effectively avoid all buildings, but also bypass potential threat areas, and finally accurately reach the target location to carry out strikes. The trajectories planned by the other five algorithms either cause the drone to accidentally collide with the building, or increase the mission risk due to improper path planning; it is particularly noteworthy that the trajectories generated by the SCA and MFO algorithms are seriously deviated, and their turning angles directly deviate from the preset three-dimensional environment, which obviously cannot achieve effective strikes on the target. Therefore, the EEFO algorithm is significantly superior to the other five algorithms in terms of safety, efficiency and accuracy.

[0145] In the scenarios of 80 urban building obstacles and 10 radar threat areas, 50 experiments were conducted on the EEFO algorithm and the other five algorithms. The average fitness results after each iteration are as follows: Figure 5-7As shown in the figure, with the deepening of the iteration process and continuous optimization, the EEFO algorithm has a higher efficiency in solving the feasible solution, which is manifested in the lower fitness value of the track and a significant downward trend. Compared with other algorithms, the EEFO algorithm can explore and lock a better flight trajectory with a higher probability, ensuring that the drone can completely avoid all potential obstacles and threat areas. After 50 iterations, the other five algorithms have a low success rate in breaking through obstacles and other constraints, and cannot find a feasible path that meets the conditions. In the complex and changing dense urban environment, especially in the face of the challenges of high-density obstacles and enemy defense areas, the EEFO algorithm can quickly find a feasible solution and is more adaptable to complex environmental scenarios.

[0146] Although the specific embodiments of the present invention have been described and illustrated in detail above, it should be noted that various changes and modifications may be made to the above embodiments without departing from the spirit of the present invention and the scope of the appended claims.

Claims

1. A method for UAV cluster target attack trajectory planning based on optimization algorithm, characterized by , including the following steps: (1) According to the characteristics of the complex urban environment, use the dimensional elevation map modeling method and MATLAB to process and obtain a three-dimensional complex urban environment model; (2) Establish the UAV constraint conditions for the complex urban environment model, and the UAV constraint conditions are the safe flight altitude area constraint, the maximum range constraint, and the waypoint obstacle avoidance area constraint; (3) Establish the UAV trajectory cost function for the complex urban environment model, evaluate the trajectories of the UAV cluster, and obtain the trajectory evaluation results; (4) Use the optimization algorithm combined with the complex urban environment model, UAV constraint conditions, and trajectory evaluation results to obtain the optimal route for the UAV cluster's trajectory planning in the complex urban environment.

2. According to the method of claim 1, the method is characterized by: , the complex urban environment model includes the modeling of urban building dangerous areas and threat areas, and the modeling of urban building dangerous areas and threat areas is constructed by using the dimensional elevation map modeling method to truly restore the information of urban building topography, threat areas, and other enemy defense measures.

3. The method for planning the target strike trajectory of a swarm of unmanned aerial vehicles based on an optimization algorithm according to claim 2 is characterized in that , the modeling of urban building dangerous areas is to set urban buildings as cylinders, and when the UAV is in this area, it is in a dangerous state. The modeling expression formula is: Among them, the parameters of the nth cylindrical city building are expressed as (a 1n ,b 1n ,h 1n ,R 1n ), (a 1n ,b 1n ) represents the coordinates of the building center area, h 1n Represents the building height, R 1n Represents the radius corresponding to the building, 1 represents the building, n is a non-zero natural number, and (x, y, z) is the coordinates of the drone track point.

4. The method for planning the target strike trajectory of a swarm of unmanned aerial vehicles based on an optimization algorithm according to claim 3 is characterized in that , the modeling of threat areas is to set threat areas as hemispheres, and the modeling expression formula is: The mth ground threat area is represented by (a 2m ,b 2m ,R 2m ), (a 2m ,b 2m ) represents the coordinates of the threat center area, 2 represents the threat point, m is a non-zero natural number, R 2m Indicates the radius corresponding to the threat point.

5. The method for planning the target strike trajectory of a swarm of unmanned aerial vehicles based on an optimization algorithm according to claim 4 is characterized in that , the safe flight altitude area constraint is: H min ≤H≤H max Among them, H represents the current flight altitude of the drone, H min is the minimum safe flight altitude, H max The maximum safe flight altitude; The maximum range constraint is: Among them, l max Indicates the maximum value of the flight range, l e Represents each flight segment, and the total flight line number of the drone is {l1,l2,l3…l e }A total of e track lines; The waypoint obstacle avoidance area constraint is: z n >A(x, y, z) or z m >B(x, y) Among them, z n represents the elevation of the UAV track point n, z m Indicates the elevation of the UAV track point m.

6. The method for planning the target attack trajectory of a swarm of unmanned aerial vehicles based on an optimization algorithm according to claim 5 is characterized in that , the evaluation of the trajectories of the UAV cluster is specifically as follows: (1) Set the performance indicators for the target strike trajectory planning of the UAV cluster: range cost, altitude change cost, and turning angle cost; (2) Through the range cost function, altitude change cost function, and turning angle cost function, obtain the trajectory cost function: J cost =D cost +H cost +Q cost Among them, J cost is the total track cost function, D cost is the flight cost function, H cost is the height variation cost function, Q cost is the turning angle cost function; The range cost function is the actual flight distance of the UAV in the UAV cluster from waypoint k to waypoint k + 1, and its expression is: The altitude change cost function is: Where C is the number of track points, is the average height of the track points, and Z is the flight altitude of the drone; The turning angle cost function is: Through the solution and processing of the trajectory cost function, obtain the trajectory composed of line segments, and then use the cubic B-spline method to smooth it, and finally obtain the actual flyable trajectory evaluation results of the UAV.

7. The method for planning the target attack trajectory of a swarm of unmanned aerial vehicles based on an optimization algorithm according to claim 6 is characterized in that , the specific steps of the optimization algorithm are as follows: S1: Automatically obtain the starting point coordinates, target point coordinates, complex urban environment model, UAV constraint conditions, trajectory evaluation results, maximum number of iterations T, and UAV population number g in the UAV trajectory planning control system; S2: Calculate the energy factor E, and the formula for calculating E is: Among them, t represents the current number of iterations, T represents the maximum number of iterations, and r is a random number within (0, 1); If E > 1, then update the UAV position according to the navigation interaction behavior, and the calculation formula is: Among them, during the navigation phase, V i (t+1) represents the position vector of UAV i at the t+1th iteration, V j (t) represents the position vector of drone j at the tth iteration, j is the position of a drone randomly selected from the current drone group and j≠i, F represents a random matrix that can randomly generate 0 and 1, represents the average position vector of the drone, and p1 represents a random number within (0,1); If 2 / 3 < E < 1, the UAV explores a new area: V i (t+1)=-r×U i (t+1)+r×H r (t+1)-L×(H r (t+1)-V i (t)) Among them, in the exploration stage, H r (t+1) represents the random position of the drone in the exploration area, r is a random number in (0,1), and L is the Levy flight function used to expand the search space; If 1 / 3 < E < 2 / 3, perform trajectory correction on the UAV, and locally optimize to approach the target path. The expression formula is: U i (t+1)=W(t)+α×|W(t)-x prey (t)| Among them, in the correction stage, U i (t+1) represents the corrected position vector of UAV i in the t+1 iteration, W(t) represents the projected position vector of the current UAV in the normalized search space, x prey (t) represents the current optimal solution position of the target UAV, α represents the initial scale of the resting zone, and controls the strength of local optimization; Update the UAV navigation position according to the correction behavior, according to the formula: V i (t+1)=U i (t+1)+n1×(U i (t+1)round(rand)×V i (t)) Among them, n1 represents a random number within (0, 1), and round(rand) represents a random function that generates 0 or 1; random function; In the attack phase, H prey (t+1) represents the position of the target UAV in the t+1th iteration, β is the reduction coefficient of the control lock area, and the scale β makes the range of the strike area smaller over time; If 0 < E < 1 / 3, update the position according to the locking behavior, and the calculation formula is: V i (t+1)=H prey (t+1)+η×(H prey (t+1)-round(rand)×V i (t)) Among them, η represents the curling factor; S3: The position of the drone is based on the fitness fit (V i (t)) to decide whether to update: Among them, fit(V i (t)) represents the current UAV’s track fitness; S4: After each iteration, update the current optimal solution x prey (t) Obtain the optimal path in the UAV trajectory.

Images