Unmanned aerial vehicle cluster surrounding control method and system for imitating eagle group cooperative hunting

Abstract

The invention discloses an eagle-group-imitating cooperative hunting unmanned aerial vehicle cluster surrounding control method and system. The method is characterized by comprising the following steps that 1, the states of unmanned aerial vehicles and a target are initialized; 2, selecting unmanned aerial vehicle cluster neighbors simulating eagle group leader rotation; 3, designing a target dynamic information observer; 4, designing a moving target surrounding control law of an imitated eagle group cooperation surrounding behavior; 5, converting an equivalent control instruction; and 6, displaying the control data surrounded by the unmanned aerial vehicle cluster. According to the method, leader rotation and cooperation surrounding mechanisms in eagle group cooperation hunting behaviors are used for reference, and under the conditions that the communication range and the link transmission capacity are limited and part of individual detection fails, through simplified neighbor selection, distributed observation and cooperation surrounding control, the reliability of the eagle group cooperation hunting behavior is improved. The dynamic adaptive capacity and control precision of the unmanned aerial vehicle cluster to the target in a complex environment are improved, and the robustness and rapidity of cooperative surrounding task execution are further improved.

Description

Technical Field

[0001] The present invention relates to a method and system for controlling the surrounding of an unmanned aerial vehicle (UAV) cluster by imitating the cooperative hunting of a flock of eagles, and belongs to the field of intelligent control and decision-making of UAV clusters. Background Art

[0002] Through the cooperative operation of multiple UAVs, the combat of UAV clusters makes up for the limitations of the individual task capabilities, and thus can undertake complex tasks that are difficult for a single UAV in the traditional sense to handle, greatly improving the task success rate. At the same time, their cooperation effectively reduces the task execution cost, showing extremely high task reliability. The cooperative surrounding task is a typical form of UAV cluster combat. By multiple UAVs cooperating to track a moving target, the target is gradually surrounded within a restricted area, providing a necessary basis for subsequent tasks such as interference and capture. Surrounding control has attracted much attention due to its high dynamics, strong confrontation characteristics, and being one of the basic combat tasks.

[0003] To control costs, due to the limitations of on-board equipment, the detection and communication capabilities of individual UAVs are limited, and there will be differences in performance, functions, etc. among different individuals, thus forming a heterogeneous cluster. Therefore, in order to achieve the cooperative tasks of a heterogeneous cluster in a complex environment, it is urgent to design an efficient cluster distributed observation mechanism, optimize the information acquisition and sharing strategy, and improve the accurate and continuous tracking ability of three-dimensional moving targets. However, most of the relevant research on cluster surrounding control focuses on two-dimensional planar moving targets, and methods such as differential game and model predictive control are used to achieve the pursuit and surrounding of stationary or moving targets. The constraints and strategies considered in the design of these methods are limited to two-dimensional scenarios and are difficult to directly extend to the complex three-dimensional space moving target scenarios suitable for UAV clusters. At this time, the target has stronger maneuverability and diverse spatial constraints, posing higher requirements for the adaptability and dynamic response of the surrounding control algorithm.

[0004] As a typical type of raptor, individual eagles usually have excellent hunting abilities. However, in the case of scarce prey resources or highly flexible prey, some eagles represented by Harris hawks exhibit various cooperative hunting behavior strategies, including leadership rotation, cooperative surrounding, etc. These strategies enable the flock of eagles to limit the target in a smaller area, significantly improving the hunting success rate, and reflecting the flexible adaptability and social cooperation ability in a complex environment. The cooperative hunting behavior of the flock of eagles provides important inspiration and technical reference for the surrounding control problem of UAV clusters in three-dimensional space, especially in task scenarios that require high efficiency and multi-aircraft cooperation, and has significant research and application value. Summary of the Invention

[0005] The present invention provides a method and system for controlling a drone swarm to surround in imitation of the cooperative hunting behavior of eagles, aiming to solve the problem of cooperative surrounding control of mobile targets by current drone swarms in a three-dimensional space. By drawing on strategies such as energy balance, leadership rotation, and cooperative surrounding in the cooperative hunting of eagles, on the basis of the cluster distributed observation mechanism, the dynamic adaptability and control accuracy of the drone swarm to the target in a complex environment are improved, and the robustness and rapidity of the execution of the cooperative surrounding task are further enhanced.

[0006] The drone swarm surrounding control system designed in the present invention to imitate the cooperative hunting behavior of eagles consists of five parts, namely: 1) neighbor selection module, 2) target observation module, 3) controller construction module, 4) motion module, and 5) data recording module, which realizes the surrounding control and demonstration of three-dimensional dynamic targets. The relationships among the various parts are as Figure 1 shown, and each part will be introduced in detail below.

[0007] 1) The neighbor selection module is a module for generating the communication interaction topology of the drone swarm for the surrounding task considering the communication ability limitations of individual drones. This module is used to implement neighbor selection based on the eagle swarm leadership rotation mechanism, and in real time, a simplified neighbor set is generated distributively for each individual, reducing the communication link burden.

[0008] 2) The target observation module is a module for collaborative observation and perception of the target state of the drone swarm considering the detection ability differences of individual drones. This target observation module includes a distributed target state observer, so that even if only some individuals have detection capabilities, the swarm can still obtain effective information about the target in real time.

[0009] 3) The controller construction module is a module for the surrounding control law of the drone swarm considering swarm cooperation. This module designs a distributed surrounding control law based on the cooperative surrounding behavior of eagles, and generates control instructions in real time according to local available information such as the target observation results and neighbor information.

[0010] 4) The drone motion module is an optional part for demonstration, including four parts: drone model, motion constraints, control instruction conversion, and target simulation, which are used to simulate individual drones to facilitate the display of the effect of the proposed surrounding control method.

[0011] 5) The data recording module is used to record the individual states, target states, observer data, etc. in the swarm, which is convenient for subsequent performance analysis and effect display.

[0012] A method for controlling a drone swarm to surround in imitation of the cooperative hunting behavior of eagles, the process is as Figure 2 shown, and the specific implementation steps are as follows:

[0013] Step 1: Initialize the states of the drones and the target

[0014] Initialize the state values such as the positions and velocities of the UAV swarm; initialize the observation values of the target, etc. For the demonstration, it is also necessary to generate UAV objects, including UAV models, motion constraints, target motion characteristics, total simulation time, and step size.

[0015] Step 2: UAV Swarm Neighbor Selection Imitating Eagle Group Leadership Rotation

[0016] Based on the leadership rotation mechanism in the cooperative hunting behavior of the eagle group, that is, the individual with higher advantages over the prey in the eagle group will take over as the new leader, the following cluster neighbor selection process is designed.

[0017] S21. Actual Neighbor Set. According to the set frequency condition, UAV i records the individuals that can receive information as neighbor individual j based on the connection status of the on-board communication device, forming the actual neighbor set at time t That is

[0018]

[0019] Among them, is the UAV set, Ξ i (j) represents the connection status judgment for determining whether UAV i can communicate with UAV j. If communication is possible, the output is 1; otherwise, the output is 0. Therefore, the individuals in the actual neighbor set are those from which the current UAV can directly obtain the required information. At this time, directly establishing communication to obtain all the required information at once will greatly occupy the communication link.

[0020] S22. Acquisition of Target Characteristics. Referring to the eagle group leadership rotation process, the UAV collects the hunting efficiency values of the initial neighbor individuals, that is, the hunting efficiency of the individual for the target, which reflects its rank in the cluster interaction topology. The hunting efficiency value δ i (t) of UAV i at time t is defined as

[0021]

[0022] Among them, δ ic (t) is the ability value of the individual for the hunting target according to its own state, δ t is the value of the prey, and b it represents the influence degree of the prey on the individual. The smaller the hunting efficiency value δ i (t), the smaller the change in the hunting success probability of UAV i for the target. Relatively speaking, it can be considered that the result is more accurate, the advantage over the target is higher, and it is more suitable to become the leader of UAV i.

[0023] S23. Neighbor Selection. Compare the hunting efficiency values of the individuals in the obtained actual neighbor set with the hunting efficiency value of UAV i itself, that is

[0024]

[0025] Among them, is the set of UAVs that can observe the target, is the set of UAVs that cannot observe the target, and there is

[0026] UAV i will establish stable communication with the individuals in the simplified neighbor set obtained by Equation (3) to obtain neighbor position, speed information, and their observation results of the target, etc., achieving the purpose of reducing communication volume.

[0027] Step 3: Design of the target dynamic information observer

[0028] Some UAV individuals no longer have the detection ability due to the lack of detection airborne equipment, equipment failure, enemy interference, etc. Therefore, the following distributed observer is designed:

[0029]

[0030] Among them, is the observed value of the target position by UAV i, p t and v t are the actual position and speed of the target, α is the observer parameter, and Φ(a, b) is a nonlinear function, defined as

[0031]

[0032] Step 4: Design of the motion target enclosing control law imitating the cooperative enclosing behavior of the eagle flock

[0033] Based on the cooperative enclosing behavior mechanism in the cooperative hunting behavior of the eagle flock, the cluster enclosing control law is designed according to the following process.

[0034] S41. Design of the potential energy function. Different roles of eagles in the cooperative enclosing process of the eagle flock will have different impacts on individual decisions. Among them, the leading eagle (leader) has a greater impact due to its higher advantage over the prey. Inspired by this, the following cluster potential function ψ 1 is designed to describe the collaborative relationship between individual i and neighbor j in the UAV cluster

[0035]

[0036] Among them, a > 0 is a parameter, d c is the expected inter-machine distance, d ij = ||p j - p i || is the distance between UAV i and neighbor j, Specifically, to simulate the influence of eagles with different roles on the individual decision-making results, a role influence function λ(i,j) is introduced, and there is

[0037]

[0038] where λ 0 is the minimum role influence coefficient.

[0039] Specifically, to simulate the degree of mutual influence among individuals at different distances within the eagle group during cooperative hunting, the following position influence activation function related to the distance between individuals is designed There is

[0040]

[0041] where h is a parameter used to adjust the interval when the output of the activation function is 1, and β 1 , β 2 is the activation adjustment parameter, and r c is the communication distance of the UAV. To effectively approach the target, the following target potential function Ψ is adopted T

[0042]

[0043] where is the distance between the UAV and the observed value of its current target position, and r T is the expected distance between the UAV and the target.

[0044] S42. Design of the UAV swarm enclosing control law. Therefore, the following UAV enclosing control law is designed

[0045]

[0046] where ω 1 , ω 2 , ω 3 , ω 4 are the controller coefficients, and u ai is the acceleration-type control command of the i-th UAV.

[0047] Step Five: Equivalent control command conversion

[0048] What is obtained in Step Four is the expected triaxial acceleration value of the UAV individual. Optionally, for system demonstration, control command conversion is used to equivalently convert the acceleration-type control command into the UAV overload-type control command, that is

[0049]

[0050] where u s =[n x , n zsinγ,n z cosγ] T is an overload-type control instruction, and n x and n z and γ are the tangential overload, normal overload, and roll angle of the UAV.

[0051] After conversion, the overload control instruction can be directly used as the input of the UAV motion model to obtain the next UAV state. When the end condition is not met, jump to step two and continue to execute; when the end condition is met, jump to step six. The end condition is determined according to requirements.

[0052] Step six: Display of UAV swarm surrounding control data

[0053] During the operation of the method proposed in the present invention, the data recording module continuously records necessary information such as the observation results of the observer, control instructions, and UAV states. Optionally, during the demonstration, the data recording module also records information such as the target state and the change in the distance between the aircraft. After the operation of the proposed method ends, the data can be retrieved to display the UAV swarm surrounding control results and demonstrate the control effect under the current settings.

[0054] The UAV swarm surrounding control method and system for imitating the cooperative hunting of eagle groups proposed by the present invention have the following advantages and effects: By drawing on the leadership rotation and cooperative surrounding mechanisms in the cooperative hunting behavior of eagle groups, under the conditions of limited communication range and link transmission capacity and partial individual detection failure, through streamlining neighbor selection, distributed observation, and cooperative surrounding control, the dynamic adaptability and control accuracy of the UAV swarm to the target in a complex environment are improved, and the robustness and rapidity of the execution of the cooperative surrounding task are further enhanced. Description of the drawings

[0055] Figure 1 is the block diagram of the UAV swarm surrounding control system for imitating the cooperative hunting of eagle groups.

[0056] Figure 2 is the flowchart of the UAV swarm surrounding control method for imitating the cooperative hunting of eagle groups.

[0057] Figure 3a and Figure 3b is the comparison diagram of the interaction before and after neighbor selection at a certain moment.

[0058] Figure 4 is the observer error diagram.

[0059] Figure 5 is the distance diagram between the UAV and the target.

[0060] Figure 6 is the position diagram of the UAV swarm and the target at the end moment (the numbers are the UAV numbers and the distances between the UAVs and the target).

[0061] Figure 7 It is a three-dimensional trajectory diagram.

[0062] The labels and symbols in the figure are explained as follows:

[0063] x, y, z - The position of the unmanned aerial vehicle (UAV) in the ground coordinate system Specific implementation manner

[0064] Next, specific examples are used to verify the effectiveness of the UAV swarm enclosing control method and system for imitating the cooperative hunting of eagle flocks proposed by the present invention.

[0065] An embodiment of the present invention, a UAV swarm enclosing control system for imitating the cooperative hunting of eagle flocks, as Figure 1 shown, consists of five parts, namely: 1) Neighbor selection module, 2) Target observation module, 3) Controller construction module, 4) Motion module, 5) Data recording module, which realizes the enclosing control and demonstration of three-dimensional dynamic targets. The following is a detailed introduction to each part.

[0066] 1) Neighbor selection module, which is a module for generating the communication interaction topology of the UAV swarm for the enclosing task considering the communication ability limitations of individual UAVs. This module is used to implement neighbor selection based on the eagle flock leadership rotation mechanism, and generate a simplified neighbor set for individuals in real time to reduce the communication link burden.

[0067] 2) Target observation module, which is a module for collaborative observation and perception of the target state of the UAV swarm considering the detection ability differences of individual UAVs. This target observation module includes a distributed target state observer, so that even if only some individuals have detection capabilities, the swarm can still obtain effective information about the target in real time.

[0068] 3) Controller construction module, which is a module for the enclosing control law of the UAV swarm considering swarm collaboration. This module designs a distributed enclosing control law based on the cooperative enclosing behavior of the eagle flock, and generates control instructions in real time according to the target observation results and neighbor information and other locally available information.

[0069] 4) UAV motion module, which is an optional part for demonstration, including four parts: UAV model, motion constraints, control instruction conversion, and target simulation, used to simulate individual UAVs to facilitate the display of the effect of the proposed enclosing control method.

[0070] 5) Data recording module, which is used to record the individual states, target states, observer data, etc. in the swarm, facilitating subsequent performance analysis and effect display.

[0071] The example is simulated and described in MATLAB 2023b, assuming that a swarm of 8 UAVs encloses a moving target.

[0072] Step 1: Initialize the UAV and the target state

[0073] Initialize the state values of the UAV swarm, such as position and velocity; initialize the observation values of the target, etc. For the demonstration, it is also necessary to generate UAV objects, including UAV models and motion constraints. In this example, the UAV model adopts a three-degree-of-freedom overload model, that is

[0074]

[0075] where p = [x, y, z] T and are the position and velocity of the UAV in the ground coordinate system, and v is the velocity magnitude; Θ = [θ, ψ, γ] T is the Euler angle of the UAV, which are the pitch angle, yaw angle, and roll angle in sequence; g is the acceleration due to gravity; n x , n z are the tangential overload and normal overload of the UAV. The overload form control instruction is defined as u s = [n x , n z sinγ, n z cosγ] T .

[0076] The motion constraints can be expressed as

[0077]

[0078] where the subscripts min and max represent the minimum and maximum allowable values of the corresponding quantities respectively. In this example, take v min = 5, v max = 70, n min = -3, n max = 3, γ min = -π / 3, γ max = π / 3, g = 9.80665.

[0079] The motion characteristics of the target are

[0080]

[0081] The total simulation duration is 70 seconds, and the time step is 0.02 seconds.

[0082] Step 2: Select neighbors of the UAV swarm by imitating the leadership rotation of the eagle flock

[0083] The process of selecting neighbors of the UAV swarm is as follows:

[0084] S21. The actual neighbor set. The actual neighbor set at time t is

[0085]

[0086] Among them, is a set of drones, Ξ i (j) is taken as

[0087]

[0088] Among them, r c is the communication distance. In this example, r c = 150.

[0089] S22. Acquisition of target characteristics. The hunting efficiency value δ i (t) of drone i at time t is defined as

[0090]

[0091] Among them, δ ic (t) is the ability value of the individual to the hunting target according to its own state, δ t is the value of the prey, and b it represents the influence degree of the prey on the individual. The smaller the hunting efficiency value δ i (t), the smaller the change in the hunting success probability of drone i to the target. Relatively speaking, it can be considered that the result is more accurate, the advantage to the target is higher, and it is more suitable to become the leader of drone i. In this example, δ ic (t) = 0.05, δ t = 1.

[0092] S23. Neighbor selection. Compare the hunting efficiency values of the individuals in the obtained actual neighbor set with the hunting efficiency value of drone i itself, that is

[0093]

[0094] Among them, is the set of drones that can observe the target, is the set of drones that cannot observe the target, and there is The interaction diagram of the drone cluster before and after neighbor selection is shown in Figure 3.

[0095] Step 3: Design of the target dynamic information observer

[0096] Some individuals no longer have the detection ability due to the lack of detection airborne equipment or due to equipment failure, enemy interference, etc., so the following distributed observer is designed:

[0097]

[0098] Among them, is the observed value of the target position by drone i, pt and v t are the actual position and velocity of the target, α is the observer gain, and in this example, α = 1.99 is taken. Φ(a, b) is a non-linear function defined as

[0099]

[0100] Step 4: Design of the motion target surrounding control law by imitating the cooperative surrounding behavior of the eagle flock

[0101] S41. Design of the potential energy function. Design the following swarm potential function ψ 1 to describe the cooperation relationship between individual i and neighbor j in the UAV swarm:

[0102]

[0103] where a > 0 is a parameter, d c is the expected inter-machine distance, d ij = ||p j - p i || is the distance between UAVs i and j, in this example, a = 4, d c = 120, σ = 0.1. The role influence function λ(i, j) is

[0104]

[0105] where λ 0 is the minimum role influence coefficient. In this example, due to the small number of individuals, λ 0 = 1 is taken.

[0106] Specifically, to simulate the mutual influence degree of individuals at different distances within the eagle flock during cooperative hunting, design the following position influence activation function related to the inter-individual distance There is

[0107]

[0108] where h is a parameter, β 1 , β 2 is the activation adjustment parameter, r c is the UAV communication distance. In this example, h = 0.1, β 1 = 1.5, β 2 = 0.5, r c = 150. Adopt the following target potential function Ψ T

[0109]

[0110] where, The distance between the drone and the observed value of its current target position, r T is the desired distance between the drone and the target. In this example, r T = 100.

[0111] S42. Design of the UAV swarm enclosing control law. Design the following UAV enclosing control law

[0112]

[0113] where ω 1 , ω 2 , ω 3 , ω 4 are the controller coefficients, and u ai is the acceleration-type control command of the i-th UAV. In this example, ω 1 = 5, ω 2 = 2, ω 3 = 6.5, ω 4 = 10.5.

[0114] Step Five: Equivalent control command conversion

[0115] What is obtained in Step Four is the desired triaxial acceleration values of the UAV individuals. Optionally, for system demonstration, control command conversion is used to equivalently convert the acceleration-type control command into the UAV overload-type control command, that is

[0116]

[0117] where u s = [n x , n z sinγ, n z cosγ] T is the overload-type control command, and n x , n z , γ are the tangential overload, normal overload and roll angle of the UAV.

[0118] After conversion, the overload control command can be directly used as the input of the UAV motion model to obtain the next UAV state. When the end condition is not met, jump to Step Two and continue to execute; when the end condition is met, jump to Step Six. The end condition is determined according to requirements.

[0119] Step Six: Display of UAV swarm enclosing control data

[0120] During the operation of the method proposed in the present invention, the data recording module continuously records necessary information such as the observation results of the observer, control commands, and the status of the UAVs. Optionally, during the demonstration, the data recording module also records information such as the target status and the change in the distance between the UAVs. After the operation of the proposed method ends, the data can be retrieved to display the control results of the UAV swarm surrounding, and demonstrate the control effect under the current settings, where the observer error is as Figure 4 shown, the change in the distance between each UAV and the target in the swarm is as Figure 5 shown, the distribution of the positions of the swarm and the target at the end of the simulation is as Figure 6 shown, and the three-dimensional motion trajectories of the swarm and the target in the simulation are as Figure 7 shown.

Claims

1. A method for controlling the surrounding of drone clusters by imitating cooperative hunting by eagle groups, characterized by: The steps of this method are as follows: Step 1: Initialize the drone and target state Step 2: Neighbor selection of drone clusters with eagle-like leader rotation Based on the leadership rotation mechanism in the cooperative hunting behavior of eagle groups, that is, the individuals with higher advantages over prey in the eagle group will take over as the new leader, a cluster neighbor selection process is designed; Step 3: Design of target dynamic information observer Distributed observers are designed for some UAVs because they do not have onboard detection equipment or no longer have detection capabilities due to equipment failure or enemy interference. Step 4: Design of moving target encirclement control law simulating the cooperative encirclement behavior of eagle flocks Based on the cooperative encirclement behavior mechanism in the cooperative hunting behavior of eagle groups, a cluster encirclement control law is designed; specifically: eagles with different roles in the process of cooperative encirclement of eagle groups will have different influences on individual decisions, among which the head eagle, i.e. the leader, has a greater influence because of its higher advantage over prey. Based on this principle, a potential energy function is designed, and a cluster encirclement control law is further designed; Step 5: Equivalent control instruction conversion Step 4 obtains the three-axis acceleration value expected by the individual drone. To demonstrate the system, control command conversion is used to convert the acceleration control command into the drone overload control command. After the conversion, the overload control command is directly used as the input of the drone motion model to obtain the next drone state. If the end condition is not met, jump to step 2 and continue execution.

2. The method according to claim 1, characterized in that: The method further includes: when the end condition is met, jumping to the following: Step 6: UAV cluster surrounds the control data display The observation results of the observer, control instructions, and information on the status of the drone are continuously recorded; during the demonstration, the data recording module will also record information on the target status and changes in the distance between drones; after the operation is completed, the data will be called up to display the results of the drone cluster encirclement control and demonstrate the control effect under the current settings.

3. The method according to claim 1 or 2, characterized in that: The process of step 2 is as follows: S21. Actual neighbor set: According to the set frequency conditions, drone i records the individuals that can receive information as neighbor individuals j according to the connection status of the airborne communication equipment, forming the actual neighbor set at time t Right now: in, For drone collection, i (j) indicates the connection status judgment of whether the UAV can communicate with the UAV. If communication is successful, the output is 1, otherwise the output is 0; S22. Target feature acquisition: Referring to the leader rotation process of the eagle group, the drone collects the hunting efficiency values ​​of the initial neighbor individuals, which describes the hunting efficiency of the individual for the target and reflects its level in the cluster interaction topology; the hunting efficiency value δ of drone i at time t i (t) is defined as: Among them, δ ic (t) is the ability value of the individual to hunt the target according to its own state, δ t For the value of the prey, and b it Indicates the degree of influence of prey on individuals; hunting efficiency value δ i The smaller (t) is, the smaller the change in the probability of successful hunting of the target by drone i is, the more accurate the result is, the higher the advantage over the target is, and the more suitable it is to become the leader of drone i; S23, neighbor selection: the actual neighbor set obtained The hunting efficiency value of the individual in is compared with the hunting efficiency value of drone i itself, that is: in, is the set of drones that can observe the target, is the set of drones that cannot observe the target, Drone i will be connected to the simplified neighbor set obtained by formula (3) Individuals in the network establish stable communication and obtain the location and speed information of their neighbors as well as their observation results of the target.

4. The method according to claim 1 or 2, characterized in that: The process of step 3 is as follows: The distributed observer: in, is the observation value of the target position by UAV i, p t and v t is the actual position and velocity of the target, α is the observer parameter, and Φ(a,b) is a nonlinear function defined as:

5. The method according to claim 1 or 2, characterized in that: The process of step 4 is as follows: S41. Potential function design: Design the following cluster potential function ψ1 to describe the collaborative relationship between individual i and neighbor j in the drone cluster: Among them, a>0 is a parameter, d c is the expected distance between machines, d ij =||p j -p i || is the distance between drone i and neighbor j, S42. Design of UAV swarm encirclement control law: Among them, ω1, ω2, ω3, ω4 are controller coefficients, u ai It is the acceleration control instruction of UAV i.

6. The method according to claim 5, characterized in that: The potential energy function design process described above includes: in order to simulate the influence of eagles of different roles on individual decision-making results, the role influence function λ(i,j) is introduced, which is: Among them, λ0 is the minimum role influence coefficient.

7. The method according to claim 5, characterized in that: The potential energy function design process, in which: in order to simulate the mutual influence degree of individuals at different distances within the hawk group during cooperative hunting, the following position influence activation function θ(x) related to the distance between individuals is designed, which is: Among them, h is a parameter used to adjust the interval when the activation function output is 1, β1, β2 are activation adjustment parameters, and r c is the communication distance of the UAV.

8. The method according to claim 5, characterized in that: The potential energy function design process described above, wherein: in order to effectively approach the target, the following target potential function Ψ is used T , in, is the distance between the drone and its current target position observation value, r T is the expected distance between the UAV and the target.

9. A drone cluster encirclement control system that simulates cooperative hunting by eagle groups, used to execute the method according to any one of claims 1 to 8, characterized in that: The system comprises: 1) The neighbor selection module is a communication interaction topology generation module for drone swarms in encirclement missions that takes into account the limitations of individual drone communication capabilities. This module is used to implement neighbor selection based on the swarm leader rotation mechanism, and to generate simplified neighbor sets for individual distribution in real time, thereby reducing the burden on communication links. 2) The target observation module is a cooperative observation and perception module for the target state of the UAV cluster that takes into account the differences in the detection capabilities of individual UAVs. The target observation module includes a distributed target state observer, so that the cluster can obtain effective information about the target in real time even if only some individuals have detection capabilities. 3) The controller building module is a UAV cluster encirclement control law module that considers cluster collaboration; this module designs a distributed encirclement control law based on the cooperative encirclement behavior of the eagle group, and generates control instructions in real time according to the target observation results and neighbor information; 4) UAV motion module, which is an optional part for demonstration, including UAV model, motion constraint, control command conversion, and target simulation. It is used to simulate individual UAVs to demonstrate the effect of the proposed encirclement control method. 5) The data recording module is used to record the individual status, target status, and observer data in the cluster to facilitate subsequent performance analysis and effect display.

Images