Manned-unmanned aerial vehicle formation assembly method, device, computer equipment and storage medium

Abstract

The present application relates to a method, device, computer equipment and storage medium for assembling a manned-unmanned aerial vehicle formation. The method includes: obtaining the flight path of the leader, the speed of the leader, the initial posture of each wingman, the speed of each wingman and the expected formation parameters of the manned-unmanned aerial vehicle formation; obtaining the position set of each wingman according to the flight path of the leader and the expected formation parameters; obtaining the current posture of each wingman according to the particle swarm algorithm; calculating the flight distance of each wingman according to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman, and calculating the target fitness function value of each wingman and the coordination fitness function value of the formation; when it is judged that the coordination fitness function value of the formation is less than or equal to a preset first threshold, outputting the Dubins curve as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman. The present method can realize the assembly process of the formation from an initial arbitrary state to the expected formation.

Description

Technical Field

[0001] The present application relates to the technical field of manned-unmanned aerial vehicle formation control, and in particular to a manned-unmanned aerial vehicle formation assembly method, apparatus, computer equipment and storage medium. Background Art

[0002] As battlefield environments and missions become increasingly complex, traditional UAV formation flight relies on ground base station control, which has limited command and control range and information transmission speed. The lack of rapid and accurate tactical decision-making in response to battlefield situation changes has greatly limited the combat effectiveness of multi-UAV formations. Manned UAV formation operations will be an important form of future air combat. Manned UAV formation operations complement the advantages of the two. Through the command and control of the entire formation by manned pilots, the pilots' wisdom and comprehensive judgment ability in the entire loop are fully utilized, which not only improves the survivability of manned aircraft, but also extends the detection and attack distance of UAVs, greatly improving the combat effectiveness of formations. In recent years, manned UAV formation technology has become a research hotspot for various military powers.

[0003] As the first step in executing a mission, the formation assembly of a manned-unmanned aerial vehicle formation is directly related to whether the formation can form the desired formation. Therefore, an effective and rapid formation assembly strategy directly affects the combat effectiveness of the formation. The formation assembly of a manned-unmanned aerial vehicle formation is a multi-aircraft trajectory planning problem with terminal constraints. It not only requires arriving at the assembly point at the same time, but also the position, attitude angle and speed of each aircraft need to meet the requirements of the desired formation. Unlike the traditional multi-unmanned aerial vehicle formation, the manned aircraft, as the leader of the formation, has the highest decision-making power and control power. Under the constraints of satisfying the time coordination and spatial coordination of each aircraft in the formation, the wingmen at different initial positions (positions and attitudes) plan their own optimal trajectories according to the flight trajectory of the leader aircraft. The final generated coordinated trajectory of the leader aircraft is the formation assembly trajectory of the manned-unmanned aerial vehicle formation.

[0004] At present, there are two methods to solve the formation assembly problem. One is to change the flight time by controlling the speed of the drones so that the drones can reach the target location at the same time; the other is to change the flight time by changing the length of the track to achieve the purpose of simultaneous arrival. Although the above methods can achieve simultaneous arrival, they are not considered to be the real completion of the formation assembly. They often ignore the motion constraint problem, and also ignore the speed and attitude constraints of the initial and terminal positions, making it difficult to ensure the assembly accuracy. Summary of the invention

[0005] Based on this, it is necessary to provide a manned-UAV formation assembly method to address the above-mentioned technical problems, which can realize the assembly process of the formation from an initial arbitrary state to the desired formation, and ensure the formation assembly accuracy and flight safety.

[0006] The manned-unmanned aerial vehicle formation assembly method includes:

[0007] Obtaining a trajectory planning task for the assembly of a manned-unmanned aerial vehicle formation; the trajectory planning task includes: the trajectory of the leader, the speed of the leader, the initial positions of each wingman, the speed of each wingman, and the expected formation parameters;

[0008] According to the flight path of the leader and the expected formation parameters, the position set of each wingman is obtained; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population, the particle state is updated, and the current position and posture of each wingman is obtained;

[0009] According to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman, based on Dubins theory, the flight distance of each wingman is calculated, and the target fitness function value of each wingman is calculated;

[0010] According to the target fitness function value of each wingman, a multi-group cooperation mechanism is introduced to calculate the collaborative fitness function value of the formation;

[0011] When it is determined that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold, a Dubins curve is output as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0012] In one of the embodiments, it further includes: when it is determined that the cooperative fitness function value of the formation is greater than a preset first threshold, updating the number of iterations, updating the current posture of each wingman according to the particle swarm algorithm, and updating the target fitness function value of each wingman based on the Dubins theory;

[0013] According to the updated target fitness function value, the collaborative fitness function value of the formation is updated and re-judged until the preset conditions are met. According to the initial posture of each wingman and the current posture of each wingman, the Dubins curve is output as the formation assembly track.

[0014] In one embodiment, after updating the target fitness function value of each wingman, the method further includes: obtaining an initial temperature, and obtaining a current temperature based on a simulated annealing algorithm according to the updated target fitness function value;

[0015] When it is determined that the current temperature is less than or equal to a preset second threshold, a Dubins curve is output as a coordinated assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0016] In one of the embodiments, it further includes: when it is determined that the current temperature is greater than a preset second threshold, updating the number of iterations, updating the current position of each wingman according to the particle swarm algorithm, and updating the target fitness function value of each wingman based on the Dubins theory;

[0017] According to the updated target fitness function value, based on the simulated annealing algorithm, the current temperature is updated and re-judged until the preset conditions are met. According to the initial posture of each wingman and the current posture of each wingman, the Dubins curve is output as the formation assembly track.

[0018] In one of the embodiments, the preset conditions include: the collaborative fitness function value of the formation is less than or equal to a preset first threshold, or the current temperature is less than or equal to a preset second threshold.

[0019] In one of the embodiments, the preset condition further includes: the number of iterations is equal to a preset third threshold.

[0020] In one embodiment, after outputting the Dubins curve as the formation assembly track, the method further includes:

[0021] According to the formation assembly track, a track tracking control law with a dual-loop structure is designed so that the position and attitude of each wingman can track the movement of the formation assembly track;

[0022] The dual loop includes an outer loop and an inner loop; the outer loop establishes a track tracking guidance law based on a vector field method; the inner loop uses the track tracking guidance law as input to design an attitude tracking control law.

[0023] A computer device comprises a memory and a processor, wherein the memory stores a computer program; a computer-readable storage medium stores the computer program; when the processor executes the computer program, and when the computer program is executed by the processor, the following steps are implemented:

[0024] Obtaining a trajectory planning task for the assembly of a manned-unmanned aerial vehicle formation; the trajectory planning task includes: the trajectory of the leader, the speed of the leader, the initial positions of each wingman, the speed of each wingman, and the expected formation parameters;

[0025] According to the flight path of the leader and the expected formation parameters, the position set of each wingman is obtained; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population, the particle state is updated, and the current position and posture of each wingman is obtained;

[0026] According to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman, based on Dubins theory, the flight distance of each wingman is calculated, and the target fitness function value of each wingman is calculated;

[0027] According to the target fitness function value of each wingman, a multi-group cooperation mechanism is introduced to calculate the collaborative fitness function value of the formation;

[0028] When it is determined that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold, a Dubins curve is output as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0029] The above-mentioned manned-unmanned aerial vehicle formation assembly method can plan the assembly trajectory of the formation, avoid collisions between aircraft in the formation, and at the same time meet the speed, heading and formation requirements between each aircraft, thereby realizing the effective assembly of the manned-unmanned aerial vehicle formation; under the Dubins-CSAPSO algorithm designed in the present invention, the maximum relative distance error of the wingman during assembly is only 0.4 meters; compared with the heuristic direct search algorithm, the Dubins-CSAPSO algorithm designed in the present invention can effectively avoid collisions between the formation and each aircraft within the formation and obstacles, thereby realizing the safe assembly of the manned-unmanned aerial vehicle formation. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 A schematic diagram of a process flow of a manned-unmanned aerial vehicle formation assembly method in one embodiment;

[0031] Figure 2 A decomposition diagram of a formation assembly control problem in one embodiment;

[0032] Figure 3 is a diagram of a formation assembly control structure in one embodiment;

[0033] Figure 4 A schematic diagram of an algorithm flow in an embodiment;

[0034] Figure 5 A schematic diagram of a wedge formation of manned and unmanned aerial vehicles in one embodiment;

[0035] Figure 6 is a diagram of a drone track tracking error model in one embodiment;

[0036] Figure 7 Schematic diagram of takeoff and assembly planning of a manned-unmanned aerial vehicle formation in one embodiment, a is a three-dimensional planning schematic diagram, and b is a two-dimensional planning schematic diagram;

[0037] Figure 8 is a graph showing a relative distance change between an MAV and a UAV in one embodiment;

[0038] Fig. 9 A diagram showing the assembly planning of formations at different positions in an embodiment;

[0039] Fig.10 is a diagram showing a comparison of assembly time differences under different algorithms in one embodiment;

[0040] Fig.11 A schematic diagram of comparing trajectory planning under different algorithms in one embodiment;

[0041] Fig.12 A diagram showing a simulation result of a UAV expected trajectory tracking in one embodiment;

[0042] Fig.13 A curve diagram showing the tracking error between the actual flight track and the expected track of the UAV in one embodiment;

[0043] Fig.14 is a control variable virtual target speed map in one embodiment;

[0044] Fig.15 is a heading rate change curve diagram in one embodiment;

[0045] Fig.16 is a graph showing a change in roll angle and heading rate in one embodiment;

[0046] Fig.17 This is a diagram showing the trajectory tracking effect of a manned-unmanned aerial vehicle formation assembly in one embodiment;

[0047] Fig.18 A partial enlarged view of the track tracking of each UAV in one embodiment;

[0048] Fig.19 It is a structural block diagram of a manned-unmanned aerial vehicle formation assembly device in one embodiment;

[0049] Fig. 20 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0051] like Figure 1 As shown, the present application provides a method for assembling a manned-unmanned aerial vehicle formation, which, in one embodiment, includes the following steps:

[0052] Step 102, obtaining the trajectory planning task of the manned-unmanned aerial vehicle formation assembly; the trajectory planning task includes: the trajectory of the leader aircraft, the speed of the leader aircraft, the initial posture of each wingman aircraft, the speed of each wingman aircraft and the expected formation parameters.

[0053] At this time, you can set the number of iterations to 1 or other values.

[0054] Step 104, according to the flight path of the leader and the expected formation parameters, the position set of each wingman is obtained; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population, the particle state is updated, and the current position and posture of each wingman is obtained.

[0055] Specifically, the flight path of the leader is discretized and combined with the expected formation parameters of the wingman to obtain the position set of each wingman. According to the particle swarm algorithm, the position set of each wingman is used as a subpopulation to update the particle state, which includes particle velocity and particle position. According to the particle state, the current posture of each wingman is obtained, which includes the current position and current heading angle.

[0056] The wingman's desired formation parameters include the priority of each wingman, so the position set of the wingman can be obtained.

[0057] Each wingman has a corresponding position set, and one wingman corresponds to a sub-population. The current position of the corresponding wingman is obtained according to the particle state in the sub-population.

[0058] Step 106, according to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman, based on Dubins theory, the flight distance of each wingman is calculated, and the target fitness function value of each wingman is calculated.

[0059] Specifically, according to the initial posture of each wingman and the current posture of each wingman, the flight distance of each wingman is calculated; according to the flight distance of each wingman, the speed of each wingman and the speed of the leader, based on the Dubins theory, the target fitness function value of each wingman is calculated.

[0060] Step 108, based on the target fitness function value of each wingman, a multi-swarm cooperation mechanism is introduced to calculate the collaborative fitness function value of the formation.

[0061] Specifically, according to the target fitness function value of each wingman, a multi-group cooperation mechanism is introduced to meet the anti-collision constraints within the formation and the obstacle avoidance constraints outside the formation, and the global optimal assembly point of each wingman is obtained; according to the global optimal assembly point of each wingman, the collaborative fitness function value of the formation is obtained.

[0062] Preferably, the constraints to be satisfied also include: satisfying boundary condition constraints, turning performance constraints, collision avoidance constraints within the formation, and obstacle avoidance constraints outside the formation.

[0063] Step 110, when it is determined that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold, a Dubins curve is output as the formation assembly track according to the initial position and current position of each wingman.

[0064] The Dubins theory is an existing technology, and the Dubins curve can be obtained according to the principle of the Dubins theory.

[0065] Preferably, it also includes: when it is determined that the collaborative fitness function value of the formation is greater than a preset first threshold, updating the number of iterations, updating the current posture of each wingman according to the particle swarm algorithm, and updating the target fitness function value of each wingman based on the Dubins theory; updating the collaborative fitness function value of the formation according to the updated target fitness function value, and re-judging until the preset conditions are met, outputting the Dubins curve as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0066] Preferably, after updating the target fitness function value of each wingman, it also includes: obtaining the initial temperature, and obtaining the current temperature based on the simulated annealing algorithm according to the updated target fitness function value; when it is judged that the current temperature is less than or equal to a preset second threshold, outputting the Dubins curve as the collaborative assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0067] Specifically, the energy change of the two iterative processes is obtained according to the target fitness function value and the updated target fitness function value; then according to the Metropolis criterion, when the energy change is less than 0, the updated value of each group position is used as the new global optimal solution, otherwise, the updated value is accepted as the global optimal solution according to the criterion probability; then the cooling operation is performed to obtain the current temperature.

[0068] Preferably, it also includes: when it is determined that the current temperature is greater than a preset second threshold, updating the number of iterations, updating the current posture of each wingman according to the particle swarm algorithm, and updating the target fitness function value of each wingman based on the Dubins theory; based on the updated target fitness function value, based on the simulated annealing algorithm, updating the current temperature, and re-judging until the preset conditions are met, outputting the Dubins curve as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0069] Preferably, the preset conditions include: the collaborative fitness function value of the formation is less than or equal to a preset first threshold, or the current temperature is less than or equal to a preset second threshold.

[0070] Preferably, the preset condition also includes: the number of iterations is equal to a preset third threshold.

[0071] The above-mentioned manned-unmanned aerial vehicle formation assembly method can plan the assembly trajectory of the formation, avoid collisions between aircraft in the formation, and at the same time meet the speed, heading and formation requirements between each aircraft, thereby realizing the effective assembly of the manned-unmanned aerial vehicle formation; under the Dubins-CSAPSO algorithm designed by the present invention, the maximum relative distance error of the wingman during assembly is only 0.4 meters; compared with the heuristic direct search algorithm, the Dubins-CSAPSO algorithm designed by the present invention can effectively avoid collisions between the formation and each aircraft within the formation and obstacles, thereby realizing the safe assembly of the manned-unmanned aerial vehicle formation.

[0072] It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows, and there is no strict order restriction. Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0073] Further preferably, when calculating the target fitness function value of each wingman and the number of calculations is greater than or equal to two times, the following judgment is performed simultaneously:

[0074] 1) According to the updated target fitness function value, the collaborative fitness function value of the formation is updated, and the collaborative fitness function value of the formation is determined;

[0075] 2) Update the current temperature and determine the current temperature according to the updated target fitness function value;

[0076] 3) Determine the number of iterations;

[0077] When the value of the formation's collaborative fitness function is less than or equal to the preset first threshold, or the current temperature is less than or equal to the preset second threshold, or the number of iterations is equal to the preset third threshold, the Dubins curve is output as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

[0078] It should be noted that the first threshold, the second threshold and the third threshold are respectively related to the collaborative fitness function value of the formation, the current temperature and the number of iterations, and can be specifically set according to actual conditions.

[0079] Further preferably, after outputting the Dubins curve as the formation assembly track, the method also includes: designing a track tracking control law of a dual-loop structure according to the formation assembly track, so that the position and attitude of each wingman tracks the movement of the formation assembly track; the dual-loop includes an outer loop and an inner loop; the outer loop establishes a track tracking guidance law based on a vector field method; the inner loop uses the track tracking guidance law as input to design an attitude tracking control law.

[0080] That is to say, after the trajectory planning of formation assembly, tracking control is also included.

[0081] This application is based on the background of manned UAV formations performing aerial missions. Aiming at the formation assembly control problem of manned UAV formations, a manned UAV formation assembly method is designed. First, considering the formation constraints, kinematic constraints and collision constraints at the head and end of the assembly, a multi-swarm cooperation mechanism is introduced into the particle swarm algorithm, and a CSAPSO formation assembly trajectory planning algorithm is designed based on the Dubins curve; then, taking the planned assembly trajectory as the expected trajectory, a nonlinear path tracking control method with a dual-loop structure is designed; finally, the effectiveness and superiority of the designed formation assembly trajectory planning method are simulated and verified using Matlab software, so that safe formation assembly can be achieved.

[0082] In a specific embodiment, the formation assembly control problem can be divided into two sub-problems: formation assembly trajectory planning and desired trajectory tracking control, such as Figure 2 As shown in the figure. The MAV / UAV formation assembly mission is assumed as follows: the lead MAV (i.e. manned aircraft) departs from the base or the air and flies according to the pre-planned track, and the wingman UAVs (i.e. unmanned aircraft) are distributed at any position in space. After the pilot issues the formation assembly command, the formation command and control system performs top-level track planning, and the planned assembly track information is transmitted to the track tracking control module. The generated track tracking control law is transmitted to the bottom control system of each UAV, thus realizing the mapping from task requirements to command issuance and then to formation flight track.

[0083] In order to realize the assembly process of the formation from any initial state to the desired formation, a MAV / UAV formation assembly control strategy framework is designed, such as Figure 3As shown in the figure, assembly control includes five levels. The first level is the formation collaborative task layer, which issues formation assembly task instructions; the second level is the central computing and processing module, which calculates the relative distance within the formation, the distance to obstacles, and the flight time of each MAV / UAV by receiving the status of each UAV and the environmental status; the third level is the trajectory planning module, which uses the designed trajectory planning algorithm to generate the optimal collaborative assembly trajectory; the fourth level is the trajectory tracking module, which uses the collaborative assembly trajectory generated by the previous level as the expected trajectory and uses the designed trajectory tracking control method to generate the attitude guidance law; the fifth level is the UAV control system, and the attitude guidance control law realizes the corresponding UAV flight action. To simplify the research problem, the following assumptions are made:

[0084] 1) Ignore the influence of aerodynamic forces on MAV and UAV and only consider the particle motion equation;

[0085] 2) The communication delay and packet loss issues within the formation are not considered;

[0086] 3) To ensure that the formation enters the formation holding phase at the same speed, it is assumed that the MAV and UAV fly at a constant speed at a predetermined altitude;

[0087] 4) When the UAVs arrive at the assembly point, the formation adopts a leader-wingman control strategy to maintain formation.

[0088] Consider the takeoff and assembly problem of N MAV / UAV formations. Suppose the mission formation contains one MAV and N-1 UAVs, and I = {l, 1, ..., N-1} is the formation sequence set.

[0089] 1. Boundary condition constraints

[0090] Taking the formation plane assembly process as an example, the initial positions of the MAV and UAV in the formation are known, and the initial state can be expressed as In order to meet the formation assembly conditions, the terminal state constraint of the wingman corresponding to the leader is:

[0091]

[0092] Where the subscripts f and l represent the states of the wingman and the leader, respectively, and Δx and Δy represent the distances between the leader and wingman in the x-direction and y-direction respectively to meet the desired formation parameters.

[0093] 2. Turning performance constraints

[0094] Fixed-wing MAVs and UAVs achieve heading deflection through rolling motion in actual flight. The turning radius R of the UAV should meet the following turning radius constraints:

[0095]

[0096] Where g is the acceleration due to gravity, φ maxTo meet the maximum roll angle required for UAV maneuverability.

[0097] 3. Anti-collision constraints within the formation

[0098] To avoid collisions between internal MAVs and UAVs during formation assembly, the following constraints must be met:

[0099]

[0100] Where x i (t), y i (t) represent the coordinates of the center of mass of the ith aircraft at time t, and L is the minimum safety distance.

[0101] 4. Obstacle avoidance constraints outside the formation

[0102] At the same time, in order to avoid collisions between the formation and obstacles, no-fly zones, etc., the following constraints must be met:

[0103]

[0104] Where x o ,y o represents the coordinates of the obstacle mass center, R o is the maximum threat radius of obstacles.

[0105] The trajectory planning problem of N MAV / UAV formation assembly is equivalent to determining an assembly position for each UAV, so that the UAV and MAV meet the speed, heading angle and formation constraints, and the overall evaluation index of all trajectories tends to be optimal or suboptimal. In order to better solve the MAV / UAV formation collaborative assembly trajectory planning problem, this section designs a parallel cooperative simulated annealing particle swarm algorithm to determine the optimal formation assembly point, and uses the Dubins curve to generate the formation assembly trajectory.

[0106] In a specific embodiment, the manned-unmanned aerial vehicle formation assembly method includes the following steps:

[0107] Step 1: Given the preset flight path γ of the MAV leader l , which is discretized as (x l (m),y l (m),ψ l (m)), m = 1, 2, ... K, the set of assembly points of each wingman UAV is calculated by given assembly formation parameters γ fi (m) = (x fi (m),y fi (m),ψ fi (m));

[0108] Step 2: Initialize each subpopulation. Set the size of each subpopulation to M and the initial temperature to T. i, end temperature T f , the current number of iterations k = 1 and the maximum number of iterations TT; initialize the speed v of all particles in each sub-population ij and position x ij , i=1,…N-1 represents the i-th subpopulation, j=1,…M represents the j-th particle in the subpopulation;

[0109] Step 3: Update the speed and position of particles in each sub-population according to the following formula (5), and calculate q using the Dubins curve theory. fi (0) and q fi (x ij ) and calculate the target fitness function value J of the i-th subpopulation according to the following equations (9), (10) and (12): i , by comparing the fitness function value, storing the local optimal p ij and the global optimal p ig ;

[0110] Step 4: Each subpopulation selects its own global optimum as a representative, and forms a group with other subpopulations in a collaborative manner, and calculates the current formation collaborative fitness function value according to the following formula (13);

[0111] Step 5: Update the target fitness function value J of each subpopulation according to the following equations (9), (10) and (12): i , get the energy change before and after the state, at the current temperature, determine the fitness value of the particle at the current temperature according to the following formula (7), and then perform the cooling operation, T = αT;

[0112] Step 6: Check the termination conditions (reaching the maximum number of iterations TT, dropping to the termination temperature T f Or the collaborative fitness function J satisfies the threshold ε). If the condition is not met, let k=k+1 and return to step 3. If the condition is met, the collaborative Dubins curve is output, and the output curve is the trajectory curve of the manned-UAV formation assembly.

[0113] Specifically, the present invention is a manned-unmanned aerial vehicle formation assembly method based on the Dubins-CSAPSO algorithm. Consider a five-aircraft formation, including one manned aircraft and four unmanned aerial vehicles. The manned aircraft is the leader of the formation, which is controlled by the pilot in real time and flies according to a preset track, and the unmanned aerial vehicles are controlled as wingmen by a control system. The flight track of the leader is known, and the assembly formation parameters are given. The assembly point position of each wingman relative to the leader is obtained by optimizing the CSAPSO algorithm, and the Dubins curve is used to generate the track of each wingman. After assembly, the desired formation is maintained until the last wingman joins the formation to complete the entire formation formation assembly task.

[0114] The Dubins-CSAPSO formation assembly trajectory planning algorithm used in the present invention has the following algorithm flow: Figure 4 shown.

[0115] The MAV manned aircraft flies as the formation leader according to the pre-designed trajectory γl. The discretization of γl can be expressed as γ l (m) = (x l (m),y l (m),ψ l (m)), where m∈{0,1,2,...,K}. According to the formation requirements, the ideal discrete point of each drone can be determined as γ fi (m) = (x fi (m),y fi (m),ψ fi (m)), where the subscript fi represents the i-th wingman drone.

[0116] like Figure 5 As shown in the figure, taking the wedge formation as an example, the MAV is the leader of the formation, and each UAV is the wingman of the MAV, and they follow the preset track of the leader to achieve the formation assembly. l Along the preset track l Flight, the track set can be discretized as q l =[x l ,y l ,ψ l ] The wingman is in the initial state q fi (0) = [x fi (0),y fi (0),ψ fi (0)] and a fixed speed v f Along the planned track fi flight, where the wingman assembly point set can be expressed as q fi =[x fi ,y fi ,ψ fi ]. Assume that the optimal assembly point q of the i-th wingman is gi Complete the formation assembly with the leader at gi , γ l and γ fi The relationship between them is as follows:

[0117] q gi =γ fi (g) = γ l (g)-d i (8)

[0118] Where d i =[Δx i ,Δy i,0] is the expected formation parameter of the i-th wingman relative to the leader, q gi =[x i (g),y i (g),ψ i (g)].

[0119] Here, the feasible solution of N-1 drones is taken as N-1 sub-populations, and each sub-population updates its speed and position in parallel according to the standard particle swarm optimization (PSO). Assume that in the D-dimensional target search space, there are M particles forming a group, and the position and speed of the jth particle can be expressed as X j =(x j1 ,x j2 ,...,x jD ) and V j =(v j1 ,v j2 ,...,v jD ), the corresponding fitness value can be determined by substituting the particle position into the fitness function. The optimal position searched by the i-th particle is called the individual extreme value p bestj =(p j1 ,p j2 ,...,p jD ), the optimal position searched by the entire particle swarm is called the global extreme value g bestj =(g j1 ,g j2 ,...,g jD ). The update formula of each particle's position and velocity is as follows:

[0120]

[0121] Where v j (k) is the current velocity of the particle; x j (k) is the current position of the particle; c 1 and c 2 is the learning factor; r 1 and r 2 is a random number uniformly distributed between [0, 1]. j is the optimal position of the particle so far, p g is the optimal position of all particles so far. λ is the compression factor, as shown in the following formula:

[0122]

[0123] The simulated annealing (SA) algorithm is a random optimization algorithm based on Monte-Carlo iterative solution. The SA algorithm starts from a relatively high initial temperature. As the temperature parameter continues to decrease, it randomly searches for the global optimal solution that optimizes the objective function through a certain probability of sudden jumps in the solution space, so as to achieve a probabilistic jump from the local optimal solution to obtain the global optimal solution. The Metropolis criterion defines the energy probability of a system changing from state i to state j at a certain temperature as:

[0124]

[0125] Where E i and E j is the energy of the solid in state i and j, E j -E i is the energy change, and K is the Boltzmann constant. The SA algorithm accepts "deteriorated" solutions with a probability. This behavior seems unreasonable, but in fact it improves the flexibility of the algorithm to a certain extent, expands the search range of the algorithm, and increases the diversity of particles. This simulated annealing strategy enables the algorithm to avoid the disadvantage of falling into the local optimal solution during the iterative search process, and improves the possibility and reliability of searching for the global optimal solution.

[0126] In order to solve the optimal assembly position of each UAV, it is particularly important to determine the fitness function of the sub-population.

[0127] In order to successfully realize the formation assembly into the desired configuration, the lead aircraft and the wingman need to arrive at the same time, that is, the flight time difference between their arrival at the assembly point is 0. In fact, the flight time difference |T e |The following relationship should be satisfied:

[0128]

[0129]

[0130] |T e |=|T f -T l |<ε (11)

[0131] Where T f and T l Respectively represent the assembly time of the wingman and the leader, v f and v l Respectively represent the flight speed of the wingman and the leader, t(.,.) represents the flight time between the initial state and the assembly state, L(.,.) represents the track length between the two states, which can be obtained by the length formula between the known two-point posture in the Dubins curve theory. ε is the threshold, q gi The determination of the assembly point is important for |Te | plays a decisive role. Here, the fitness function of the i-th drone is defined as:

[0132] J i =|T fi -T l | (12)

[0133] In order to obtain a safe and optimal formation assembly trajectory, this section regards the formation as a population, and each drone in the formation as a subpopulation. In the entire large population, in order to avoid collisions between subpopulations and collisions with external obstacles, each subpopulation will perform a cooperative evaluation between subpopulations after calculating each adaptive function, and share the individual extreme values ​​obtained by the current subpopulations to form a new group. By correcting the collaborative fitness function of the formation population, updating the individual extreme value and the group extreme value, it is determined whether the optimal value obtained by each subpopulation meets the formation assembly constraints, so as to find the optimal (or suboptimal) value that satisfies the formation collaboration.

[0134] In order to plan a safe and flyable formation assembly trajectory, it is necessary to consider the collision constraints within the formation and the obstacle avoidance constraints outside the formation. Assuming that each aircraft in the formation can obtain each other's position information and obstacle information in real time, the condition for no collision within the formation is to satisfy the collision avoidance constraints within the formation of formula (3), and the condition for no collision with obstacles, no-fly zones, etc. outside the formation is to satisfy the obstacle avoidance constraints outside the formation of formula (4).

[0135] The anti-collision constraint and obstacle avoidance constraint are added to the fitness function as a penalty function, and the formation coordination fitness function is defined as follows:

[0136]

[0137] Where k inner and k outer They represent the penalty factors for anti-collision inside and outside the formation, and H is the number of obstacles. When a collision occurs inside or outside the formation, J 1 or J 2 If it is greater than 0, the penalty term plays a constraint role; when the formation avoids obstacles or no collision occurs within the formation, J 1 or J 2 is equal to 0. At this time, the optimal group extreme value of each particle group is the assembly point position of each drone in the formation that can safely fly the trajectory.

[0138] When there is a collision between MAV / UAVs in the formation or between UAVs and obstacles during the assembly process, the cooperative fitness function J is obviously a large number, which does not meet the accuracy of the cooperative fitness function. The population will re-update each sub-population. In order to improve the efficiency of the algorithm optimization and avoid the two UAVs that collided changing their tracks at the same time, a priority is set for the formation MAV / UAVs, that is, the aircraft with high priority can maintain the original planned track, and the aircraft with low priority will be re-planned based on the collision. The original assembly point does not meet the safety assembly requirements and needs to be recalculated to meet the safety assembly point. In the entire MAV / UAV formation, MAV has the highest level, and all UAVs obey the pilot's command. Secondly, the UAV level is determined according to factors such as UAV type and fuel volume. UAVs with heavy tasks have a high level, and UAVs with relatively less fuel have a high level. In this way, UAV1, UAV2, UAV3 and UAV4 are determined in turn.

[0139] Under the condition that the initial postures of manned and unmanned aircraft and the postures of the assembly point are known, the Dubins curve can be used to directly generate the formation assembly trajectory curve. The obtained trajectory curve is the optimal assembly trajectory of the formation, which not only meets the formation assembly accuracy but also ensures the safety of formation flight.

[0140] In order to achieve accurate tracking of the desired trajectory by the UAV, the design of the path tracking controller is to design the trajectory tracking guidance law and attitude tracking control law for any given geometric route, including straight lines, circles and curves, so that the UAV position and attitude can track the desired trajectory and continue to follow the desired trajectory. In the process of tracking general curves, the UAV may encounter singularity problems under certain circumstances, that is, tracking failure or spin. To avoid singularity, this section takes a single UAV as the research object and the planned assembly trajectory as the desired trajectory. The kinematic equation of the trajectory tracking error is:

[0141]

[0142] Where e x ,e y represent the longitudinal and lateral deviation errors, respectively, d represents the heading angle of the desired track, κ(s) is the curvature of the virtual target on the desired track, is the difference between the heading angle of the UAV and the heading angle of the desired track, v and ω represent the airspeed and heading rate control instructions of the UAV respectively, represent the wind speed and azimuth of the steady wind field, respectively. Indicates the speed of the virtual target on the expected track. Figure 6 shown.

[0143] The track tracking control module is an important part of the UAV control system. A dual-loop UAV nonlinear track tracking controller is designed. The outer loop uses the virtual target speed and heading rate as control variables, and designs the UAV track tracking guidance law based on the vector field method; the inner loop uses the outer loop guidance law as the control input, and designs the roll angle control law to track the desired heading angle of the outer loop, thereby achieving the desired track tracking.

[0144] In order to guide the UAV to the desired track, an asymptotically stable vector field is constructed near the desired track as follows:

[0145] χ d (e y )=-χ ∞ tanh(k d e y ) (15)

[0146] In the formula, k d >0 is the gain parameter, used to control e y Convergence speed; χ ∞ ∈(0,π / 2] is a constant angle, for any e y satisfy:

[0147]

[0148] When the side offset error and the azimuth error are both equal to 0, the desired track is tracked. In the constructed vector field, the desired heading angle of the UAV is given by χ d (e y ), it can be seen from formula (15) that when the UAV is far away from the expected track, the azimuth error of the UAV is χ ∞ Tracking the expected track, the UAV moves in a direction perpendicular to the expected path; as the side offset gradually decreases, the UAV gradually deflects until it is parallel to the tangent direction of the expected track.

[0149] In order to design the track tracking control law, the Lyapunov function is defined according to the track tracking error model (14) as follows:

[0150]

[0151] By taking the derivative of formula (16), we can get:

[0152]

[0153]

[0154] Set the virtual target speed according to formula (17): The control law of the UAV yaw rate ω is:

[0155]

[0156]

[0157] Substituting equations (18) and (19) into equation (17), we can obtain

[0158]

[0159] When e y ≤0, 0≤χ d (e y )≤π / 2, then e y vsinχ d (e y )≤0; when e y ≥0, -π / 2≤χ d (e y )≤0, then e y vsinχ d (e y )≤0. So for any e y , always satisfying e y vsinχ d (e y )≤0, that is, V 1 ≥0; In summary, the track tracking error model (14) is globally asymptotically stable under the control laws (18) and (19), that is, Asymptotically approaches 0.

[0160] The inner loop attitude tracking controller is designed based on the nonlinear method to make the UAV's roll angle track the desired heading rate obtained by the outer loop. When the UAV performs a horizontal turning maneuver, it usually needs to use the roll attitude control system to adjust the ailerons to tilt the UAV to generate side force to achieve the UAV's heading deflection, which can be expressed as:

[0161]

[0162] Where: φ is the UAV roll angle. Considering the UAV control input constraints, the UAV roll angle satisfies the following conditions: -φ max ≤φ≤φ max

[0163] To design the roll angle tracking control law, we introduce the auxiliary control input Combining equations (14) and (21), the track tracking error system is obtained as follows:

[0164]

[0165] Where: p is the roll angular velocity, u is the roll angle control variable.

[0166] Combined with formula (21), the roll angle state error is defined as follows (23), and the derivative of formula (23) is obtained:

[0167]

[0168]

[0169] Where: γ=V / g; ω d Obtained by formula (19).

[0170] In order to obtain a suitable roll angle control law u, the Lyapunov function is designed as follows:

[0171]

[0172] Where: λ>0, q>1 are optional parameters.

[0173] V 2 The derivative is:

[0174]

[0175] The designed roll angle control law is:

[0176]

[0177] Where: k p ,k c is a positive real number. Substituting formula (27) into formula (26) yields:

[0178]

[0179] When the parameters meet and Can be achieved

[0180] In summary, the design of control law (27) can achieve V 2 ≥0, That is, to realize the state variable e x →0,e y →0, and φ→φ d .

[0181] In order to verify the effectiveness of the formation assembly control strategy of this application, Matlab simulations were performed on the trajectory planning algorithm and trajectory tracking control method. The simulation experiment environment is Inter Core i7, 2.30GHz processor, Windows 10 operating system, and the simulation software is MATLAB R2016b.

[0182] When conducting the trajectory planning algorithm simulation experiment, the basic parameters of the simulation algorithm are shown in Table 1. This section assumes two formation assembly situations: the takeoff assembly of the maritime formation and the assembly of the air formation with different initial states. Experiment 1 is that there are no obstacles in the takeoff assembly airspace, and the MAV / UAV ascent and the aircraft carrier deck support factors are considered, and finally a wedge formation is formed in the assembly area. Experiment 2 is that MAV / UAV completes the formation assembly in different initial states in the airspace with a no-fly zone to form a wedge formation. The simulation parameters of Experiments 1 and 2 are shown in Tables 2 and 3.

[0183] Table 1 Basic simulation parameters

[0184]

[0185]

[0186] Table 2 Simulation parameters of Experiment 1

[0187]

[0188] Table 3 Experiment 2 simulation initial position parameters

[0189]

[0190] Figure 7 The three-dimensional and two-dimensional trajectory planning diagrams of Experiment 1. As shown in the figure, the MAV first takes off from the aircraft carrier to the assembly altitude and flies according to the preset trajectory. UAV1-4 waits for the aircraft carrier deck support and then takes off to the assembly altitude in turn. At the same time, they complete the assembly with MAV in turn and continue to fly in formation. When UAV4 and MAV arrive at the assembly point, the entire formation completes the entire assembly process safely.

[0191] like Figure 8 The following is a graph showing the distance between MAV and UAV1-4 changing over time. The time when UAV1 reaches the assembly altitude is used as the starting point, and the distance between MAV and UAV changes over time. At 91s, UAV1 arrives at the assembly point, and the distance between MAV and UAV1 is the expected distance of the formation, and is maintained. During this process, the minimum distance between MAV and UAV1 is 54m, which exceeds the anti-collision distance within the formation. Similarly, at 137s, 184s and 234.5s, UAV2, UAV3 and UAV4 arrive at the assembly point respectively, that is, the relative distance is the expected distance. Considering the take-off time, deck support time and level flight assembly time, the assembly time of the entire formation is 303s.

[0192] like Fig. 9This is the trajectory planning diagram for Experiment 2. In the initial stage, the MAV and the four UAVs are in different positions. At this time, the MAV pilot issues a formation assembly command, and the four UAVs plan the assembly trajectory according to the preset flight trajectory of the MAV. In the figure, the formation can avoid the no-fly zone in the algorithm designed in this paper, and the aircraft in the formation always maintain a safe distance, that is, the formation achieves a safe formation assembly, and the entire formation assembly time is 126.5s. Combining the trajectories planned in the two simulations, it can be seen that the algorithm designed in this paper can effectively plan an assembly trajectory for the entire formation, which not only avoids collisions between MAV / UAVs in the formation, but also prevents collisions between UAVs and obstacles. At the same time, it meets the speed, heading and formation requirements between MAVs and UAVs in the formation, and achieves safe and effective assembly.

[0193] In order to verify the superiority of the trajectory planning algorithm designed in this paper, the algorithm in this paper is compared with the cooperative genetic algorithm (CGA) for analysis. Table 4 below shows the assembly time difference between each UAV and MAV under the algorithm in this paper and the cooperative genetic algorithm. The algorithm designed in this paper meets the time threshold requirement. Since the UAV flies very fast, the maximum error is only 0.4m, which can be ignored. However, under the CGA algorithm, although the time is also very small, the maximum relative distance between the MAV and the UAV reaches 48m, and the error is large. The formation cannot form a formation and cannot be assembled. Fig.10 shown.

[0194] Table 4 The assembly time difference between UAV and MAV

[0195]

[0196] like Fig.11 This is a comparison chart of the algorithm in this paper and the heuristic direct search algorithm. It can be clearly seen from the figure that the planned trajectory generated by UAV1-3 is the same as that of the algorithm in this paper when the heuristic direct search algorithm is used to determine the assembly point, but UAV4 crosses the no-fly zone and cannot fly safely, that is, this planning fails. The time used by this application is significantly faster than the CGA algorithm and the heuristic direct search algorithm, which greatly improves the planning efficiency.

[0197] When conducting the trajectory tracking control simulation experiment, the effectiveness of the designed trajectory tracking control algorithm is first verified. The straight line, circle and Clothoid curve are used as the expected trajectory. The simulation time is 110s, the expected flight speed of the UAV is v = 20m / s, and the initial position of the UAV is p 1 =[50,100], the initial heading angle is The initial roll angle is φ 1 =0°, the wind speed in the steady wind field is v w =4m / s, wind field azimuth The control parameter is chosen as k d=0.1, k ω =0.1, k x =0.5, χ ∞ =π / 2, k p =0.5, k c =0.1, the control input constraints of UAV are ω∈[-0.49rad / s,0.49rad / s] and φ∈[-π / 4rad,π / 4rad]. The simulation results are shown in the figure.

[0198] from Fig.12 and Fig.13 It can be seen that the path tracking control law designed in this paper can accurately track the expected track curve under the conditions of steady wind field and UAV kinematics, and the tracking model error is kept within 0.2m; Fig.14 It can be seen that the desired track speed converges to the UAV speed setting value of 20m / s under the action of the controller; Fig.15 The heading rate changes significantly in the initial stage of tracking the expected curve. The expected track is tracked at the maximum and minimum heading rates according to the initial position and heading angle of the UAV, and converges to the expected value after 12.5s. Fig.16 The inner loop roll angle tracking result, the UAV determines the heading rate by the roll angle, and vice versa, the roll angle change curve can be obtained by tracking the heading rate. In the initial stage, the UAV roll angle changes rapidly from 0° to φ min , and then the attitude tracking of the roll angle is achieved under the action of the control law.

[0199] Taking the formation assembly trajectory planned in Experiment 2 as the expected trajectory, the practicality of the UAV nonlinear trajectory tracking controller was simulated and verified.

[0200] In the simulation, the flight speed of MAV and UAV is 200m / s, the wind speed of the steady wind field is, and the azimuth of the wind field is. Compared with the initial posture of each UAV in the trajectory planning part, this section assumes that there is an initial error in the initial state of each UAV. The specific initial state parameters of MAV and UAV are shown in Table 5.

[0201] Table 5 Initial pose parameters

[0202]

[0203] The effect of MAV / UAV formation assembly track tracking is shown in the figure below Figure 17-18 As shown. The black solid line in the figure represents the planned formation assembly track, and the red solid line is the actual tracking track. Under the action of the track tracking controller, each UAV can accurately track the expected track. However, due to the high flight speed of the drone, it can be seen from the figure that the two curves overlap. In order to make a clearer and more intuitive comparison between the expected track and the tracking track, Fig.18 The local enlarged pictures of each UAV in the initial stage are given. Fig.17It can be seen that in the presence of initial errors, each UAV asymptotically tends to the desired track and tracks the desired track accurately and stably, realizing the process from the pilot issuing the assembly task to the assembly track planning everywhere, and then to the assembly track tracking control, verifying the practicality of the nonlinear track tracking controller.

[0204] In one embodiment, Fig.19 As shown, a manned-unmanned aerial vehicle formation assembly device is provided, including: an acquisition module 1902, an update module 1904, a target fitness function value calculation module 1906, a collaborative fitness function value calculation module 1908 and a collaborative fitness function value calculation module 1910, wherein:

[0205] The acquisition module 1902 is used to acquire the trajectory planning task of the manned-unmanned aerial vehicle formation assembly; the trajectory planning task includes: the trajectory of the leader, the speed of the leader, the initial posture of each wingman, the speed of each wingman, and the expected formation parameters of the wingman;

[0206] The updating module 1904 is used to obtain the position set of each wingman according to the flight path of the leader and the expected formation parameters of the wingman; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population, the particle state is updated, and the current posture of each wingman is obtained;

[0207] The target fitness function value calculation module 1906 is used to calculate the flight distance of each wingman and the target fitness function value of each wingman according to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman based on the Dubins curve theory;

[0208] A collaborative fitness function value calculation module 1908 is used to introduce a multi-group cooperation mechanism according to the target fitness function value of each wingman to calculate the collaborative fitness function value of the formation;

[0209] The output module 1910 is used to output a Dubins curve as a formation assembly track according to the initial posture of each wingman and the current posture of each wingman when it is determined that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold.

[0210] For the specific definition of the manned-unmanned aerial vehicle formation assembly device, please refer to the definition of the manned-unmanned aerial vehicle formation assembly method above, which will not be repeated here. Each module in the above device can be implemented in whole or in part by software, hardware and a combination thereof. The above modules can be embedded in or independent of the processor in the computer device in the form of hardware, or stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0211] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Fig. 20 As shown. The computer device includes a processor, a memory, a network interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for assembling a manned unmanned aerial vehicle formation is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse. The computer device can be a simulation device, the input device inputs relevant information to the simulation device, the processor executes the program in the memory to perform combined simulation, and the display screen displays relevant simulation results.

[0212] Those skilled in the art will understand that Fig. 20 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0213] In one embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method in the above embodiment when executing the computer program.

[0214] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of the method in the above embodiment are implemented.

[0215] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0216] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0217] The above-mentioned embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the invention patent. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the attached claims.

Claims

1. A method for assembling a manned-unmanned aerial vehicle formation, characterized in that: include: Obtain the trajectory planning task for the assembly of manned and unmanned aerial vehicle formations; The trajectory planning task includes: the leader's trajectory, the leader's speed, the initial positions of each wingman, the speed of each wingman, and the expected formation parameters; According to the flight path of the leader and the expected formation parameters, the position set of each wingman is obtained; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population, the particle state is updated, and the current position and posture of each wingman is obtained; According to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman, based on Dubins theory, the flight distance of each wingman is calculated, and the target fitness function value of each wingman is calculated; According to the target fitness function value of each wingman, a multi-swarm cooperation mechanism is introduced to calculate the coordinated fitness function value of the formation; According to the target fitness function value of each wingman, a multi-swarm cooperation mechanism is introduced to calculate the coordinated fitness function value of the formation, specifically including: according to the target fitness function value of each wingman, a multi-swarm cooperation mechanism is introduced to meet the anti-collision constraint within the formation and the obstacle avoidance constraint outside the formation, and obtain the global optimal assembly point of each wingman; According to the global optimal assembly point of each wingman, the coordinated fitness function value of the formation is obtained; When it is determined that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold, a Dubins curve is output as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman.

2. The method according to claim 1, characterized in that Also includes: When it is determined that the cooperative fitness function value of the formation is greater than the preset first threshold, the number of iterations is updated, the current position of each wingman is updated according to the particle swarm algorithm, and the target fitness function value of each wingman is updated based on the Dubins theory; According to the updated target fitness function value, the collaborative fitness function value of the formation is updated and re-judged until the preset conditions are met. According to the initial posture of each wingman and the current posture of each wingman, the Dubins curve is output as the formation assembly track.

3. The method according to claim 2, characterized in that After updating the target fitness function value of each wingman, the method further includes: obtaining an initial temperature, and obtaining a current temperature based on a simulated annealing algorithm according to the updated target fitness function value; When it is determined that the current temperature is less than or equal to a preset second threshold, a Dubins curve is output as a coordinated assembly track according to the initial posture of each wingman and the current posture of each wingman.

4. The method according to claim 3, characterized in that Also includes: When it is determined that the current temperature is greater than the preset second threshold, the number of iterations is updated, the current position of each wingman is updated according to the particle swarm algorithm, and the target fitness function value of each wingman is updated based on the Dubins theory; According to the updated target fitness function value, based on the simulated annealing algorithm, the current temperature is updated and re-judged until the preset conditions are met. According to the initial posture of each wingman and the current posture of each wingman, the Dubins curve is output as the formation assembly track.

5. The method according to any one of claims 2 to 4, characterized in that: The preset conditions include: the collaborative fitness function value of the formation is less than or equal to a preset first threshold, or the current temperature is less than or equal to a preset second threshold.

6. The method according to claim 5, characterized in that The preset condition also includes: the number of iterations is equal to a preset third threshold.

7. The method according to any one of claims 1 to 4, characterized in that: After outputting the Dubins curve as the formation assembly track, it also includes: According to the formation assembly track, a track tracking control law with a dual-loop structure is designed so that the position and attitude of each wingman can track the movement of the formation assembly track; The dual loop includes an outer loop and an inner loop; the outer loop establishes a track tracking guidance law based on a vector field method; the inner loop uses the track tracking guidance law as input to design an attitude tracking control law.

8. A manned-unmanned aerial vehicle formation assembly device, characterized in that: include: The acquisition module is used to obtain the trajectory planning task of the manned-UAV formation assembly; The trajectory planning task includes: the leader's trajectory, the leader's speed, the initial positions of each wingman, the speed of each wingman, and the expected formation parameters of the wingman; The update module is used to obtain the position set of each wingman according to the flight path of the leader and the expected formation parameters of the wingman; according to the particle swarm algorithm, the position set of each wingman is used as a sub-population to update the particle state and obtain the current position and posture of each wingman; The target fitness function value calculation module is used to calculate the flight distance of each wingman and the target fitness function value of each wingman according to the speed of the leader, the initial posture of each wingman, the speed of each wingman and the current posture of each wingman based on the Dubins curve theory; The collaborative fitness function value calculation module is used to introduce a multi-population cooperation mechanism according to the target fitness function value of each wingman, and calculate the collaborative fitness function value of the formation; according to the target fitness function value of each wingman, introduce a multi-population cooperation mechanism, and calculate the collaborative fitness function value of the formation, specifically including: according to the target fitness function value of each wingman, introduce a multi-population cooperation mechanism to meet the anti-collision constraint within the formation and the obstacle avoidance constraint outside the formation, and obtain the global optimal assembly point of each wingman; according to the global optimal assembly point of each wingman, obtain the collaborative fitness function value of the formation; The output module is used to output the Dubins curve as the formation assembly track according to the initial posture of each wingman and the current posture of each wingman when judging that the value of the formation's collaborative fitness function is less than or equal to a preset first threshold.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images