Group mixed hunting control method based on space-time alternating hierarchical leadership structure

Abstract

The application relates to a kind of group mixed hunting control methods based on space-time alternative hierarchical leadership structure, belong to the technical field of robot group.This application combines the space-time alternative hierarchical leadership structure of biological group behavior, informed leadership, mixed hunting control strategy and other bionics principles, optimizes the task execution efficiency and robustness under space-time dynamic environment.The method applies the space-time alternative hierarchical leadership structure characteristics existing in biological group behavior to the control method of robot group, optimizes the collaborative behavior of robot group, proposes the bionic group mixed hunting control method, so as to realize the more robust, efficient and flexible collaboration of robot group in dynamic hunting process.

Description

Technical Field

[0001] This invention belongs to the field of robot swarm technology, specifically relating to a swarm hybrid encirclement control method based on a spatiotemporally alternating hierarchical leadership structure. Background Technology

[0002] With the rapid development of robot swarm technology, the need for dynamic encirclement tasks in heterogeneous dynamic environments (such as complex and unknown terrain, communication barriers, incomplete information, and uncertain target movement) is becoming increasingly urgent. However, existing control technologies still mainly rely on centralized global scheduling or distributed autonomous decision-making, making it difficult to achieve a dynamic balance between robustness and global optimization capabilities in scenarios with incomplete information, communication barriers, and complex terrain. Centralized control relies on a central node for global decision-making and task allocation, such as the machine learning-based swarm control method in patent CN118642377B, which optimizes multi-robot collaboration through a centralized algorithm. However, such methods have extremely high requirements for communication bandwidth and stability, and their flexibility and robustness are insufficient. In dynamic encirclement environments with communication barriers, they are prone to system collapse due to single-point failures. Distributed control emphasizes individual autonomous decision-making and the collaborative effect between robots, such as the distributed multi-robot synchronous swarm control algorithm in patent CN109164826A, which achieves the connectivity of the network topology during swarm evolution through local information interaction. While such methods possess robustness, their adaptive capabilities are limited in complex environments with incomplete information, they lack global coordination capabilities, and they suffer from efficiency bottlenecks in dynamic task allocation. Although hybrid control methods can theoretically combine the advantages of distributed and centralized approaches, there is currently a lack of relevant patented technologies for complex ground environments, and research and applications of biomimetic-based hybrid control methods are scarce.

[0003] Faced with complex and ever-changing environments, biological groups (such as sheep flocks) exhibit a unique spatiotemporal hierarchical leadership structure to improve decision-making efficiency and cope with emergencies. Group members dynamically switch between leader and follower roles through random processes, forming a time-dimensional hierarchical decision-making mechanism that ensures both local autonomous response and global coordination. This biomimetic feature reveals the inherent rationality of hybrid trapping control strategies, namely, the fusion of centralized global optimization and distributed autonomous decision-making, balancing system robustness and task efficiency. However, existing technologies lack sufficient biomimetic applications of biological behavior and exhibit weak spatiotemporal dynamic adaptability; they also demonstrate poor coordination and difficulty in collaborative exploration when facing unknown environments. The patent field still lacks hybrid trapping control methods for unstructured scenarios such as complex and unknown terrain, dynamic trapping, incomplete information, and limited communication. Summary of the Invention

[0004] (a) Technical problems to be solved

[0005] The technical problem to be solved by this invention is: how to optimize the collaborative behavior of robot groups to achieve more robust, efficient and flexible collaboration of robot groups in the dynamic capture process.

[0006] (II) Technical Solution

[0007] To address the aforementioned technical problems, this invention provides a group-based hybrid encirclement and control method based on a spatiotemporally alternating hierarchical leadership structure, comprising the following steps:

[0008] Based on the characteristics of biological swarm encirclement behavior and spatiotemporal alternating hierarchical leadership structure, the dynamic encirclement process of robot swarms is divided into different stages (TP) on the time scale, namely, three stages: search (TP1), tracking (TP2), and encirclement (TP3). For these three stages, according to the needs of robot swarm collaboration, a hybrid encirclement control strategy combining centralized and distributed approaches is adopted in the spatial dimension to achieve hybrid encirclement control of the target.

[0009] (III) Beneficial Effects

[0010] The present invention has the following advantages:

[0011] (1) This invention draws on the behavioral patterns of biological groups in nature. Based on the biological behavior and movement characteristics of groups, the group encirclement process is divided into multiple stages, including search, tracking, and encirclement stages. Each stage has its specific group behavior control mode and task objectives, enabling the robot group to complete the encirclement task in a more orderly manner. This division method allows the robot group to allocate and adjust tasks more specifically, which will effectively improve the collaborative efficiency of the robot group and the success rate of target encirclement.

[0012] (2) This invention innovatively proposes a hybrid encirclement control method, which deeply integrates the advantages of centralized and distributed control, enhances the flexibility of dynamically adjusting control strategies according to task requirements, overcomes the defects of single-mode control in dynamic encirclement processes, optimizes resource allocation and energy consumption, and improves encirclement efficiency. Centralized control realizes the overall coordination and task allocation of the robot group, providing global information to ensure the orderly conduct of the encirclement task; under distributed control, each robot can make autonomous decisions and actions based on its own perceived local information, giving individual autonomy and enhancing the system's flexibility and adaptability. In the group encirclement process, the hybrid encirclement control method achieves group collaboration through information sharing and local decision-making, which can more effectively and flexibly respond to complex terrain conditions with communication obstruction and complex and ever-changing encirclement environments, thereby improving the system's reliability and fault tolerance.

[0013] (3) This invention combines the spatiotemporal alternating hierarchical leadership structure in group biological behavior with the dynamic process of robot group encirclement. By constructing a hierarchical leadership structure, different control methods are adopted according to the needs of dynamic encirclement based on the spatial distribution of the group at different movement time stages. The upper-level centralized control method generates the global encirclement path and task priority; the lower-level distributed control method perceives the environment in real time and dynamically switches the role of "temporary leader" to trigger local encirclement strategy adjustments. This invention breaks through the problems of task continuity and response lag in dynamic target tracking under communication interruption scenarios.

[0014] (4) The present invention adopts the leadership design based on informed strategy in the leadership behavior of group animals, which enables the leader to formulate a more accurate and reasonable encirclement strategy based on his own experience, combined with the information held by informed followers and their perception of the environment. The leader can flexibly adjust the robot's movement trajectory and speed according to the actual situation, give full play to the leader's decision-making ability and the information advantage of group members, improve the comprehensiveness and accuracy of robot group decision-making, and ensure the efficient and smooth completion of the encirclement task. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the overall process of robot swarm capture in this invention.

[0016] Figure 2 This is a flowchart of the robot swarm search phase of the present invention;

[0017] Figure 3 This is a schematic diagram of the leader selection scheme for the tracking phase of the present invention;

[0018] Figure 4 This is a schematic diagram illustrating the control and decision-making principle of the robot swarm encirclement phase of the present invention.

[0019] Figure 5 This is a schematic diagram of the robot's perception area. Detailed Implementation

[0020] To make the objectives, contents, and advantages of the present invention clearer, the specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples.

[0021] This invention combines biomimetic principles such as the spatiotemporal alternating hierarchical leadership structure, informed leadership, and hybrid encirclement control strategies of biological groups to optimize task execution efficiency and robustness in dynamic spatiotemporal environments. This method applies the spatiotemporal alternating hierarchical leadership structure characteristics of biological groups to the control method of robot swarms, optimizing the collaborative behavior of robot swarms and proposing a biomimetic hybrid encirclement control method, thereby achieving more robust, efficient, and flexible collaboration of robot swarms in dynamic encirclement processes.

[0022] This invention provides a hybrid encirclement control method for robot swarms based on a spatiotemporally alternating hierarchical leadership structure, comprising the following steps:

[0023] A recent study by GómezNava et al. discovered a spatiotemporally alternating hierarchical leadership structure in which the group exhibits a leader on different timescales at different movement stages, guiding the group's movement. Experimental models revealed that the group randomly alternates between leader and follower roles (intermittent collective strategies). Based on the characteristics of biological swarm encirclement behavior and spatiotemporally alternating hierarchical leadership structures, the dynamic encirclement process of robot swarms is divided into different phases (TP) on a timescale, such as... Figure 1 As shown, the process includes three phases: search (TP1), tracking (TP2), and encirclement (TP3). For these three phases, a hybrid encirclement control strategy combining centralized and distributed approaches is adopted in the spatial dimension to achieve hybrid encirclement control of the target, based on the need for group coordination.

[0024] During the search phase, the robot swarm employs a distributed control strategy to search for and capture the target, as follows: Figure 2 As shown, all robots receive environmental information, acquire environmental parameters, form a terrain model, and autonomously search for the target. When a robot finds the target, the entire robot group enters the tracking phase. If no target is found, the search continues.

[0025] During the tracking phase, the robot swarm employs a centralized strategy to track the target. The leader selection method has two modes (either one is sufficient): leader selection based on visual information (leader_visual) and leader selection based on distance information (leader_distance). The specific process is as follows: Figure 3 As shown.

[0026] (1) The process of selecting the leader_distance based on distance information is as follows: At a certain moment, after the robot discovers the target, it synchronizes the target location information to all robots, calculates the distance xi_target between each robot and the target, calculates the minimum value min(xi_target) of the distance between all robots and the target, and defines the robot closest to the target in the robot group as the leader_distance.

[0027] (2) The process of selecting leader_visual based on visual information is as follows: at a certain moment, record the time point ti_target when the robot detects the target, calculate the minimum value min(ti_target) for all robot detection time points, and define the robot in the robot group that first detects the target in visual information as leader_visual.

[0028] After the leader (leader_visual or leader_distance) is determined, the leader commands the robot swarm to begin the tracking process based on the direction of the target. During the tracking process, the leader calculates the distance xi_target between itself and the target. When xi_target is less than or equal to the preset capture distance threshold_distance, the leader initiates a capture command, and the robot swarm enters the capture phase; otherwise, the tracking continues.

[0029] During the encirclement phase, the robot swarm employs a hybrid encirclement control strategy combining centralized and distributed approaches to capture the target. In this phase, the robot swarm is divided into 'a' groups (number of members in each group = total number of robots / number of groups 'a'). Each group operates under a distributed control strategy. This distributed control strategy has no central node; all robots are autonomous, possessing a certain degree of decision-making ability and capable of rapidly responding to environmental changes without relying on a central control point. The leader in this phase employs an informed leadership strategy. If each group contains an informed follower (a robot capable of perceiving local information (environment, direction, target, etc.), this informed follower transmits the perceived local information to the leader. The leader then combines this information with the informed follower's data to make global directional decisions and real-time assessments of the encirclement status. A control decision-making diagram is shown below. Figure 4 As shown. When determining the encirclement status, the leader needs to judge whether the target has been successfully encircled based on the degree of closure of the encirclement. If the encirclement is successful, the robot group enters the next round of target search phase; if not, the encirclement continues.

[0030] In all three stages above, the robot needs to calculate the expected direction. The specific calculation methods for each stage are introduced below.

[0031] Let i be the index of a robot in a swarm, and n represent the number of individual robots in the swarm, i = 1, 2, 3, ..., n. At each time point t, the position vector corresponding to robot i is x. i (t), the direction vector is v i (t), the maintained velocity is s iAt each time step, the robot determines the position of its neighboring robots within the three sensory regions (attraction region Az, repulsion region Rz, and orientation region Oz) based on their positions (e.g., ...). Figure 5 (As shown) to calculate the expected direction (desired direction) d i (t). A swarm interaction network describes the topological structure and dynamic interaction rules of information transmission among individuals in a swarm of robots. The social connectivity matrix is ​​a mathematical abstraction of the swarm interaction network, typically quantifying the connection strength or interaction weights between individual robots in matrix form. Therefore, the pairwise connection strength between robots, S, is predefined. ij (t) represents the group interaction network, denoted by the social connection matrix between individuals, S ij The value of (t) represents the relative influence of robot j on robot i. If robot i can be influenced by robot j, then S ij (t)≠0, otherwise S ij (t) = 0. The structure of the swarm interaction network can be fixed or configured to change over time during different stages of robot swarm movement.

[0032]

[0033] Where, for any i, j, S ij When all t are greater than 0, the robot swarm is in the search phase, and the swarm interaction network is a fully connected network; when S ij When the (t) part is greater than 0, the robot group is in the tracking or encirclement stage. The group interaction network is a partially connected network, as shown in formula (1).

[0034] To avoid collisions between individual robots, an attraction zone Az, a repulsion zone Rz, and a direction zone Oz are defined for each robot. Within the attraction zone Az, each robot adjusts its orientation to align with that of adjacent robots to prevent the group from scattering. Within the repulsion zone Rz, each robot actively moves away from adjacent robots to avoid collisions. Within the direction zone Oz, each robot adjusts its orientation to align with that of the group of robots.

[0035] The first region is the exclusion zone Rz, which maintains the robot's "individual space". If a robot j appears in robot i's Rz region, at each time step Δt (representing the direction of motion at the next moment), robot i calculates the expected direction d based on the relationships with neighboring robots in the three perception regions. i (t+Δt|TP), where TP corresponds to the position and velocity at different stages. The specific calculation formula is based on the classic couzin aggregation motion model, incorporating a group interaction network s. ij (t) and noise ε i(t) is used to reflect the changes in the connection characteristics between groups at different motion stages and the impact of noise brought about during the motion process. Therefore, in this invention, the expected direction of robot i for the next step is defined as follows: (2)

[0036]

[0037] Where, ε i (t) represents the random noise of robot i, ε i (t)~N(0,σ) 2 By adding random noise in the direction, the influence of noise from complex terrain on the direction can be simulated, which can improve the robustness of the system.

[0038] This anticipated direction ensures that no collisions will occur during aggregation within t+Δt. To ensure that no robots collide, robot j should approach other robots and enter their attraction zone Az, with its direction facing the orientation zone Oz.

[0039] When using a distributed control strategy to search for and capture targets during the search phase, it is necessary to calculate the expected direction and adjust the direction for the next moment based on the calculated expected direction. During the search phase, the attraction intensity α1 of the attraction region Az and the motion direction of the direction region Oz will affect the expected direction of robot i. At this time, the expected direction for the next step is defined as follows: (3)

[0040]

[0041] Where α1 is the regional attraction intensity of the attraction region Az during the search phase, α1∈(0,1).

[0042] During the tracking phase, a centralized control strategy is employed to track the target. The robot swarm is organized according to a structural leader, with the leader determining the swarm's direction of movement. k is defined as... i Let k be the level of an individual within the robot swarm. i =1 represents the leader, k i =2 represents the choice of followers and leaders, such as... Figure 3 As shown. Furthermore, based on whether the leader remains constant during the tracking phase and whether they change dynamically over time, the methods of leader retention are divided into the following three categories:

[0043] (1) If the leader is fixed in this stage, then k i fixed;

[0044] (2) If the leader is randomly fixed in this stage, then k i (t) changes over time. At time t, a robot is randomly selected from the group and defined as the leader.

[0045] (3) During this stage, the leader's fixed method varies with the time and space dimensions.

[0046] During the tracking phase, the attraction intensity α2 in the attraction region Az and the motion direction in the direction region Oz both affect the expected direction of robot i. Enhancing group aggregation during the tracking phase can increase the attraction intensity, where α2 > α1. At this point, the leader robot i(k) i =1) The expected direction for the next step is defined as shown in the following formula (4):

[0047]

[0048] Normalization:

[0049]

[0050] Because the individual robot's movement direction needs to be adjusted according to the leader's direction during the tracking phase, the leader's direction is dominant at this stage. In the robot swarm, besides the leader, the expected direction of other robots must consider not only the presence of neighboring robots within their perception range but also the leader's movement direction. The preference strength for the leader's movement direction is denoted by W. The larger the value of W, the greater the influence of the leader's movement direction on the next movement direction of robot i. At this point, the follower robot i(k) i =2) The expected direction for the next step is defined as shown in the following formula (6):

[0051]

[0052] Where W is the strength of the follower robot's preference for the leader's direction of movement, W>1. k＝1 (t) represents the velocity of the leader robot at time t;

[0053] In the encirclement phase, a hybrid encirclement control strategy is used to encircle the target. The robot group is organized according to informed leadership, consisting of a leader and followers. The direction of group movement is dominated by the leader and is influenced by the real-time local perception information of other robots. The leader is fixed in the same way as in the tracking phase, with three ways (fixed leader, random leader, and leader changes with time and space). Further enhancing the group aggregation in the encirclement phase can increase the attraction intensity. Let the attraction intensity of the area Az in the encirclement phase be α3, and α3>α2. The expected direction of robot i in the next step is defined as shown in the following formula (7):

[0054]

[0055] Normalizing formula (7) yields formula (8):

[0056]

[0057] To simulate the informed leader model, a preference direction g based on "knowledge" is defined for some followers. i The leader integrates the preferred directions of informed followers to aid decision-making and achieve more efficient decision-making. Since the expected direction is jointly influenced by the directions of both the leader and informed followers, the expected directions of the leader and informed followers are weighted according to a weight w. Here, w determines the degree of participation of informed leader information in directional decision-making; the smaller w is, the higher the degree of participation. In this encirclement phase, the leader robot i(k) i =1, the expected direction of the next step for the leader robot is shown in the following formula (9):

[0058]

[0059] Here, w represents the strength of the leader's preference for the direction of informed followers, where w ∈ (0,1). w = 0 indicates that the leader is completely...

[0060] When the leader relies on the preferences of informed followers, w=1 represents the leader completely relying on their own preferences and not being influenced by the preferences of informed followers. When w is between 0 and 1, it represents the leader partially incorporating the preferences of informed followers into their decision-making. In this case, the follower robot i(k) i =2) The expected direction of the next step is defined as shown in the following formula (10), where the meaning of W is the same as in formula (6);

[0061]

[0062] Criteria for successful encirclement: The closure degree of the encirclement is used as an indicator to measure the integrity of the encirclement in a robot group encirclement task. The higher the closure degree, the more complete the encirclement, and the lower the possibility of the target escaping. The formula for the closure degree of the encirclement (11) is shown below. Let Q be the closure degree of the encirclement, ρ be the actual perimeter of the encirclement, and LC be the perimeter of the ideal encirclement that is completely closed. Then the closure degree Q can be expressed as:

[0063]

[0064] Formula (12) assumes that the perimeter of the ideal encirclement is known, which can be the perimeter of a perfect circle formed around the target. Assuming that R is the radius of the encirclement, the calculation of LC is as shown in Formula (12):

[0065] LC=2πR (12)

[0066] A threshold Q_threshold is set for the closed boundary of the robot group. The leader determines that when all robots form a closed boundary, that is, when the closure degree Q is greater than or equal to the defined closed boundary threshold Q_threshold (Q≥Q_threshold), the target is considered to be completely surrounded inside, and the encirclement is successful.

[0067] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A group-based hybrid encirclement and control method based on a spatiotemporally alternating hierarchical leadership structure, characterized in that, Includes the following steps: Based on the characteristics of biological swarm encirclement behavior and spatiotemporal alternating hierarchical leadership structure, the dynamic encirclement process of robot swarms is divided into different stages (TP) on a time scale, namely, three stages: search (TP1), tracking (TP2), and encirclement (TP3). For these three stages, in the search stage, the robot swarm adopts a distributed control strategy to search for the target to be encircled; in the tracking stage, the robot swarm adopts a centralized strategy to track the target to be encircled. During the encirclement phase, the robot swarm employs a hybrid encirclement control strategy combining centralized and distributed approaches to capture the target. The swarm is divided into 'a' groups, each operating under a distributed control strategy. In this distributed control phase, all robots are autonomous, possessing a certain degree of decision-making ability and capable of responding autonomously to environmental changes. The leader in this phase adopts an informed leadership strategy; if each group has an informed follower, the follower can transmit locally perceived information to the leader. The leader then combines this information with global directional decisions and real-time assessments of the encirclement status. The informed follower is a robot capable of perceiving locally perceived information. When assessing the encirclement status, the leader determines whether the target has been successfully captured based on the closure of the encirclement. If successful, the swarm enters the next target search phase; otherwise, the encirclement continues. Let the sequence number of a robot in the robot swarm be... , where n represents the number of individual robots in the robot swarm. =1,2,3……n; at each time point A certain robot The corresponding position vector is The direction vector is The speed at which it is maintained is In each time step, the robot calculates the expected orientation based on the positions of neighboring robots in the three perception regions: the attraction region Az, the repulsion region Rz, and the orientation region Oz. This paper defines a swarm interaction network to describe the topological structure and dynamic interaction rules of information transmission between individuals in a swarm of robots. The social connectivity matrix is ​​a mathematical abstraction of the swarm interaction network, quantifying the connection strength or interaction weights between individual robots in matrix form. The connection strength between each pair of robots is predefined. For group interaction networks, it is represented by a social connection matrix among individuals. The value represents the robot j For robots i The relative influence, if the robot i Can be used by robots j Impact, then ,otherwise =0; In different stages of robot swarm movement, the structure of the swarm interaction network is fixed or set to change over time; During the encirclement phase, the robot swarm organizes itself according to informed leadership, consisting of a leader and followers. The direction of swarm movement is dominated by the leader and influenced by real-time local perception information from other robots. This further enhances swarm aggregation and increases the regional attraction intensity of the encirclement zone Az. , Let Az be the attraction intensity in the attraction zone during the tracking phase. The expected direction of robot i for the next step is defined by formula (7): （7） in, To eliminate noise; further, normalizing formula (7) yields formula (8): （8） During the encirclement phase, in order to simulate the strategy of informed leadership, a knowledge-based preference direction is defined for some followers. Leaders integrate the preferences of informed followers to aid decision-making; since expected directions are influenced by both the leader's and informed followers' directions, the expected directions of the leader and informed followers are weighted according to a weight w. This determines the degree to which informed leaders participate in directional decision-making. The smaller the value, the higher the level of participation. The expected direction of the leader robot i in this encirclement phase is updated to formula (9): （9） in, It represents the strength of a leader's preference for the direction of informed followers. ; =0 indicates that the leader completely relies on the preferences of informed followers. =1 represents that the leader relies entirely on their own preferences and is not influenced by the preferences of informed followers. Between 0 and 1, it represents the leader making decisions based on the preferences of informed followers; at this point, the expected direction of follower i for the next step is updated to formula (10): （10）。 2. The method as described in claim 1, characterized in that, During the search phase, all robots in the swarm receive environmental information, acquire environmental parameters, and form a terrain model to autonomously search for the target. When a robot finds the target, all robots in the swarm enter the tracking phase. If no target is found, all robots continue the search.

3. The method as described in claim 2, characterized in that, During the tracking phase, one of the following two modes is used to select the leader: leader_visual based on visual information and leader_distance based on distance information. The process of selecting the leader_distance based on distance information is as follows: At a certain moment, after the robot discovers the target, it synchronizes the target's location information to all robots, calculates the distance xi_target between each robot and the target, calculates the minimum value min(xi_target) of the distance between all robots and the target, and defines the robot closest to the target in the robot group as the leader_distance. The process of selecting leader_visual based on visual information is as follows: at a certain moment, record the time point ti_target when the robot detects the target, calculate the minimum value min(ti_target) for all robot detection time points, and define the robot in the robot group that first detects the target in visual information as leader_visual. Once a leader is identified, the leader commands other robots to begin tracking the target based on its direction. During the tracking process, the leader calculates the distance xi_target between itself and the target. When xi_target is less than or equal to the preset capture distance threshold_distance, the leader initiates a capture command, and the robot swarm enters the capture phase; otherwise, the tracking continues.

4. The method as described in claim 1, characterized in that, For any robot i , j , When all values ​​are greater than 0, the robot swarm is in the search phase, and the swarm interaction network is a fully connected network; when... When a portion of the value is greater than 0, the robot swarm is in the tracking or encirclement phase, and the swarm interaction network is a partially connected network. Within the attraction zone Az, individuals adjust their orientation to align with that of adjacent robots to prevent the robot group from scattering; within the repulsion zone Rz, individuals actively move away from adjacent robots to avoid collisions; within the orientation zone Oz, individuals adjust their orientation to align with that of the robot group. At different stages If a robot j appears within the repulsion region Rz of robot i, for each time step Robot i calculates the expected direction based on the relationships between itself and neighboring robots within its three perception areas. The method for calculating the expected direction is to introduce a group interaction network based on the couzin aggregation motion model. and noise To reflect the changes in the connection characteristics between groups at different motion stages and the impact of noise during the motion process, the expected direction of robot i for the next step is defined as formula (2): （2） in, Let the random noise be that of robot i. .

5. The method as described in claim 4, characterized in that, During the search phase, a distributed control strategy is employed to calculate the expected direction when searching for and capturing the target. The direction is then adjusted for the next moment based on the calculated expected direction. The attraction intensity in the attraction zone Az during the search phase is also considered. The direction of motion of both the direction region Oz and the direction region Oz will affect the expected direction of robot i. At this time, the expected direction of the next step is defined by formula (3): （3） in To determine the regional attraction intensity of attraction zone Az during the search phase, .

6. The method as described in claim 5, characterized in that, During the tracking phase, the robot swarm is organized according to a structural leader, with the leader determining the swarm's direction of movement and defining... This refers to the rank of an individual within a robot swarm, where... =1 represents the leader. =2 represents followers. Based on whether the leader remains fixed during the tracking phase and whether it changes dynamically over time, the methods of leader fixation are divided into the following three cases: (1) If the leader is fixed during this tracking phase, then fixed; (2) If the leader is randomly fixed during this tracking phase, then It changes over time, at a certain point in time. A robot is randomly selected from the group and defined as the leader; (3) During this tracking phase, the leader's fixed method varies with the time and space dimensions; Regional attraction intensity in the attraction zone Az during the tracking phase The direction of movement in both the direction region Oz and the direction region Oz will affect the expected direction of robot i, thus enhancing group aggregation during the tracking phase. At this point, the expected direction of the leader robot i for the next step is defined by formula (4): （4） Then normalized to: （5） In a robot swarm, besides the leader, the expected direction of other robots should not only take into account the situation of neighboring robots within the perception range, but also the movement direction of the leader. Let W be the preference intensity for the leader's movement direction. The larger the value of W, the greater the influence of the leader's movement direction on the next movement direction of robot i. At this time, the expected direction of the next movement of follower i is defined by formula (6): （6） in, It is the strength of followers' preference for the leader's direction of movement. , Indicates the leader robot at a certain point in time. The speed.

7. The method as described in claim 6, characterized in that, When judging the encirclement status, the criteria for determining whether the encirclement was successful are as follows: the closure degree of the encirclement is used as an indicator to measure the integrity of the encirclement in the robot group encirclement task. The higher the closure degree, the more complete the encirclement is, and the lower the possibility of the captured target escaping. Let Q be the closure degree of the encirclement. L The actual perimeter of the encirclement. Let Q be the perimeter of a perfectly closed ideal enclosing loop. Then the enclosing loop closure degree Q is expressed as: （11） Assuming the perimeter of the ideal encirclement is known, it is the circumference of a perfect circle formed around the target. Let R be the radius of the encirclement. The calculation formula is: （12） Set a threshold Q_threshold for the closed boundary of the robot group. The leader determines that when all robots form a closed boundary, that is, when the closure degree Q of the encirclement is greater than or equal to the defined closed boundary threshold Q_threshold, the target is considered to be completely surrounded inside, and the encirclement is considered successful.

Images