Robot cluster formation and cooperative hunting method based on target allocation

Abstract

The invention discloses a robot cluster formation and cooperative hunting method based on target allocation, and belongs to the technical field of robot control. Aiming at the target formation of different geometric configurations, constructing a differential initial target distribution strategy, and carrying out dynamic motion control on the robot based on a differential model until the target formation is realized; a global target redistribution mechanism, an intra-neighborhood global information sharing mechanism and a target exchange mechanism are introduced into motion control, formation convergence delay caused by the fact that individual robots are too far away from target points is effectively eliminated, and formation construction time is remarkably shortened. On the basis of the cluster formation method, a collaborative hunting method is derived, target points are dynamically generated by designing a clustering rule, hunting of targets is achieved, hunting formation can be adaptively switched, flexibility and environmental adaptability of hunting tasks in a complex environment are remarkably enhanced, and the method is suitable for large-scale popularization and application. When the formation pattern or the escaper position set is changed, the method can be quickly migrated and applied.

Description

Technical Field

[0001] This invention relates to the field of robot control technology, specifically to a method for robot swarm formation and cooperative capture based on target allocation. Background Technology

[0002] In recent years, robot swarming and encirclement technologies have attracted much attention due to their enormous application prospects in fields such as search and rescue, environmental monitoring, and intelligent logistics. Swarming aims to form a predetermined geometric configuration through the cooperative control of multiple robots to perform complex collaborative tasks; encirclement technology focuses on tracking, surrounding, or capturing targets within a specific area, with typical applications including search and rescue of trapped personnel in disaster scenarios and monitoring and dispersing wild animals in ecological protection. Given their broad application value, efficient and robust swarming and encirclement algorithms have become a key research direction for research institutions both domestically and internationally.

[0003] However, existing swarm formation algorithms still face many challenges in practical applications. On the one hand, traditional algorithms typically employ a uniform control strategy, which can lead to individual robots being too far from the target point at their initial position, significantly prolonging the overall formation time and reducing formation efficiency. On the other hand, existing encirclement algorithms mostly rely on fixed encirclement formations, making it difficult to adaptively adjust according to the movement state of the escaping target, thus affecting the effectiveness and durability of the encirclement effect. These adaptability and robustness issues are particularly prominent in dynamic environments. Summary of the Invention

[0004] This invention aims to address the aforementioned technical deficiencies by proposing a target allocation-based robot swarm formation and cooperative encirclement method. It optimizes the swarm target selection strategy to improve formation speed, shorten the total swarm movement distance, reduce energy consumption, and dynamically adjust the encirclement formation based on target escape behavior, achieving flexible and efficient encirclement.

[0005] The specific technical solution of this invention is as follows:

[0006] In a first aspect, the present invention proposes a robot swarm formation method based on target allocation, comprising:

[0007] Each robot is assigned an initial target index based on the target formation of the cluster formation, and the robot is dynamically controlled based on the differential speed model until the target formation is achieved; a global target reassignment mechanism, a neighborhood global information sharing mechanism, and a target exchange mechanism are introduced in the process of achieving the target formation.

[0008] The global target redistribution mechanism refers to the following: when the target formation is a closed polygon with adjacent target points distributed at equal intervals and the proportion of robots that have reached the target points exceeds the first threshold, robots that have not reached the target points and whose distance exceeds the second threshold are recorded as lagging robots. The nearest target point is selected for the lagging robots, and the selected target point is used as the starting point and the nearest unoccupied target point to the starting point is used as the ending point. The target points of the robots on the path from the starting point to the ending point are redistributed and locked in sequence to achieve target redistribution and accelerate formation convergence.

[0009] The neighborhood global information sharing mechanism refers to sharing global information within the neighborhood through robot neighborhood communication, marking unassigned target points discovered by the robot as candidate target points, and updating the robot gradient value to dynamically assign candidate target points;

[0010] The target exchange mechanism refers to iteratively optimizing the target points of each robot within the neighborhood, with the goal of minimizing the total formation time and total path distance.

[0011] Furthermore, an initial target index is assigned to each robot based on the target formation of the cluster, including:

[0012] When the target formation is a polygonal combination with the center line parallel to the coordinate axis, it is first clustered and grouped according to the robot's coordinate position, and then the center allocation strategy within the group or the random allocation strategy is executed.

[0013] The remaining target formations will be assigned to a central distribution strategy;

[0014] The specific center allocation strategy is as follows: first, calculate the center position of the robot cluster, and then allocate the nearest idle robot to each target point in order of distance from the target point to the cluster center position.

[0015] Furthermore, based on the differential speed model, dynamic motion control of the robot is performed until the target formation is achieved, including:

[0016] The robot's guiding velocity component and collision avoidance velocity component are calculated and synthesized, and then weighted with the velocity of the previous moment to obtain the current combined velocity;

[0017] The resultant velocity is converted into the left and right wheel velocities. A limiting constraint on the wheel velocities is introduced. The linear velocity and angular velocity of the robot are solved by the forward kinematics model, and the robot's position and attitude are updated.

[0018] Furthermore, in the global target redistribution mechanism, once the robot reaches the redistributed target point, it unlocks, and the target point in the locked state does not participate in the target exchange mechanism.

[0019] Secondly, the present invention provides a robot cooperative encirclement method based on the above-mentioned robot swarm formation method, comprising the following steps:

[0020] (1) Based on the distance matrix between the pursuers and the escapees, a balanced number of pursuers are assigned to each escapee. The correspondence between the pursuers and the escapees remains unchanged throughout the entire pursuit process.

[0021] (2) Dynamic target point generation: Dynamic clustering and grouping are performed based on the distance matrix of the escapee's location. The center and radius of each cluster are calculated. The cluster center position is adjusted through global iterative optimization to avoid overlapping of the encirclement circles. Target points equal to the number of encirclers are generated in the circumference of the encirclement circle.

[0022] (3) Initial allocation of target points: If the cluster exists in the historical group, the historical target point allocation scheme is reused; otherwise, the distance priority principle is used to allocate initial target points to the hunters within the cluster. When target points conflict, the target point is allocated to the hunter farthest from the target point, and the remaining hunters select the nearest target point from the remaining target points.

[0023] (4) Motion control: Using the above-mentioned robot swarm formation method, based on the global target redistribution mechanism, the neighborhood global information sharing mechanism and the target exchange mechanism, the hunters corresponding to each encirclement circle are controlled to move towards the target point to form the target formation of the encirclement circle;

[0024] (5) Real-time judgment of the encirclement status. If the encirclement is completed, the process ends; otherwise, return to step (2) at each time the target location is updated, until the encirclement is completed.

[0025] Furthermore, the process of globally iteratively optimizing and adjusting the cluster center positions includes:

[0026] Detect the center and radius of the new cluster and the existing clusters, and determine whether the capture circle of the new cluster overlaps with the capture circle of the existing clusters. If so, translate the center of the new cluster along the vector direction of the line connecting the centers of the conflicting clusters until the capture circles of all clusters no longer overlap.

[0027] Furthermore, the dynamic clustering process also includes:

[0028] Determine whether the current cluster is included in the cluster grouping of the previous time step. If so, and the changes in the cluster center and radius are both less than the threshold, then use the cluster center and radius of this cluster from the previous round of cluster grouping. Otherwise, use the currently calculated cluster center and radius.

[0029] Thirdly, the present invention provides a robot cooperative trapping system for implementing the above-mentioned robot cooperative trapping method.

[0030] Fourthly, the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described target-assignment-based robot swarm formation method.

[0031] Fifthly, the present invention provides a computer electronic device, including a memory and a processor;

[0032] The memory is used to store computer programs;

[0033] The processor is configured to implement the target-assignment-based robot swarm formation method described above when executing the computer program.

[0034] The beneficial effects of this invention are:

[0035] (1) In order to solve the problem that the formation of the formation is slow due to a few robots failing to reach the target position, this invention constructs a global target redistribution mechanism, optimizes the cluster target selection strategy in real time, effectively shortens the time required for formation and improves the overall system efficiency; at the same time, it introduces a global information sharing mechanism and a target exchange mechanism in the neighborhood to optimize the dynamic motion control logic.

[0036] (2) Based on the target spatial distribution characteristics of escapees, this invention designs a dynamic target point generation method based on clustering rules. This method can achieve adaptive switching of the encirclement formation while achieving the encirclement target, significantly enhancing the flexibility and environmental adaptability of the encirclement task in complex environments.

[0037] (3) The present invention has good portability. When the formation or the set of escapee positions changes, only the corresponding target position parameters need to be adjusted to realize the rapid migration and application of the method, avoiding large-scale redesign and debugging. Attached Figure Description

[0038] Figure 1 This is the overall flowchart of the cluster formation method.

[0039] Figure 2 This is a schematic diagram illustrating the principle of the center allocation algorithm for initial target allocation.

[0040] Figure 3 It is a flowchart of the complete operation of a single robot in one cycle.

[0041] Figure 4 This is the overall flowchart of the collaborative encirclement method.

[0042] Figure 5 This is the flowchart of the initial allocation algorithm for the encirclement and capture group.

[0043] Figure 6 This is a flowchart of the algorithm for calculating the clustering of escapees and the set of target points for capture.

[0044] Figure 7 This is a flowchart of the target assignment algorithm within clustering.

[0045] Figure 8 This is a schematic diagram of the process of cluster formation into the "ZJU" formation.

[0046] Figure 9 This is a diagram illustrating the process of switching from the "ZJU" formation to the "ROBOT" formation.

[0047] Figure 10 This is a schematic diagram illustrating the process of switching from a "circular" formation to an "eagle" formation.

[0048] Figure 11 This is a schematic diagram illustrating the clustering and merging process of escapees due to changes in their location during the capture operation.

[0049] Figure 12 This is a schematic diagram illustrating the clustering and decomposition process of escapees during the capture operation, due to changes in their location.

[0050] Figure 13 This is a schematic diagram of the process from "initialization of the location of the escapee / capturer" to "completion of the coordinated capture mission".

[0051] Figure 14 This diagram illustrates how, after the capture mission was completed, the location of the escapee changed, and the pursuers reorganized and successfully completed the capture. Detailed Implementation

[0052] The cluster formation and cooperative encirclement functions of the present invention will be further described in detail below with reference to specific embodiments. Those skilled in the art should understand that the embodiments described below are for illustrative purposes only and should not be considered as limiting the scope of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0053] This invention proposes a target allocation-based robot swarm formation and cooperative encirclement method. It aims to optimize the selection of target points for robot swarms through strategies such as differentiated initial target allocation methods and global target reassignment mechanisms, thereby improving formation speed, reducing the total movement distance of the swarm during formation, reducing energy loss, and dynamically generating a set of encirclement target points based on the real-time location distribution characteristics of escapees using clustering ideas to achieve adaptive reconstruction of the encirclement formation.

[0054] like Figure 1 As shown, a robot swarm formation method based on target assignment sequentially completes a series of preset target formations. For each target formation, the following steps are adopted:

[0055] S1: Based on the target formation of the cluster, select a suitable initial target allocation scheme for the robot cluster, so that each robot is assigned an initial target index.

[0056] Specifically, if the target formation is a configuration composed of multiple simple polygons on a two-dimensional plane, and the line connecting the centers of each polygon is parallel to the X-axis or Y-axis, then an initial coarse target allocation is performed based on the X / Y coordinates of each center and the X / Y coordinates of the robots in the cluster. Robots corresponding to the number of target points are selected for each polygon, and then the final target index is determined within each polygon using a random allocation or center allocation strategy. For example... Figure 2 As shown, the pentagram formation composed of circles represents the target formation, with each circle representing a target point. The triangular array represents the robot cluster, with each triangle representing a robot. The center allocation strategy is implemented as follows: the position center of the robot cluster is calculated, and then the nearest idle robot is selected for pairing with each target point in order from farthest to nearest, in order to minimize the variance of the time it takes for each robot within the cluster to reach its target point and accelerate the cluster formation speed.

[0057] This invention addresses the shape characteristics of target formations by employing a differentiated initial target allocation mechanism. By minimizing the initial distance between the robot and the target point, it effectively avoids generating robot-target point combinations with excessively large distances, thus establishing an optimized foundation for subsequent rapid formation.

[0058] S2: Dynamic motion control of the robot based on the differential speed model.

[0059] In this step, a two-dimensional differential speed model is used as the simulation model to update the position and attitude of the robots in the cluster at the current time step. Specifically, the guiding component and collision avoidance component of each robot's velocity are calculated, synthesized, and weighted with the velocity at the previous moment to obtain a new resultant velocity. The left and right wheel velocities are calculated and limited based on the current orientation angle and the constructed differential speed inverse kinematics model. Finally, the linear velocity and angular velocity are calculated using the forward kinematics model to update the cluster's position and orientation angle, thus updating the robot position and attitude. The derivation of the forward and inverse kinematics models of the differential speed model used is as follows:

[0060] Positive kinematics:

[0061]

[0062]

[0063] Inverse kinematics:

[0064]

[0065]

[0066] in, , These are the wheel speeds of the robot's left and right wheels, respectively. , These are the robot's linear velocity and angular velocity, respectively. The wheelbase is the distance between the left and right wheels. The vertical distance from the robot's center to the line connecting the left and right wheels is given. Therefore, this model is universally applicable to any differential speed model; only modifications to the algorithm are needed. and The algorithm migration can be completed by setting the value.

[0067] S3: Initiate global target reallocation for lagging robots.

[0068] Check whether the cluster meets the global adjustment target allocation judgment condition. If it does, reconstruct the target allocation. Set a distance threshold to determine whether the robot has reached the target point. If the distance between the two is less than the threshold, it is determined that it has been reached. For example, in formula (5) of this invention, the judgment threshold is set to 0.01.

[0069]

[0070] When the target formation is a simple polygon and the proportion of robots reaching the target point exceeds a threshold (i.e., the majority of robots), in order to avoid slow formation due to a few robots still being too far from their target point, it is necessary to change their target to the nearest target point A. Then, based on the characteristic that adjacent target points of a simple polygon are equidistant, a clockwise / counterclockwise search is performed until the nearest unoccupied target point B is found. The target points of the robots on the path are then redistributed sequentially in the direction from A to B to achieve the effect of overall movement, thereby greatly reducing the formation time of the small formation.

[0071] In this embodiment, a closed polygon with equidistant adjacent target points is defined as a simple polygon, and the following is adopted: This indicates that the proportion of robots that have reached the target location exceeds a set threshold (preferably 0.9, but can be adjusted as needed). When the global adjustment trigger condition is met, a global target reallocation is performed, assigning the nearest spatial neighbor target point to robots that are too far from the target point. Then, the target is updated sequentially according to the found direction. In addition, the robot index involved in the global target update needs to be set with a target locking mechanism to prevent it from being reselected as the target in subsequent target exchange stages because it is too close to the currently occupied target point. Locked target points cannot be modified by exchange rules; the lock is only released when the robot reaches the new target point, and the target exchange permission is reopened.

[0072] S4: Global information sharing within the neighborhood is achieved through robot neighborhood communication and gradient algorithms. Unassigned target points are marked as candidate target points, allowing them to be selected as targets as quickly as possible through global information sharing. Simultaneously, robot gradients within the neighborhood are analyzed and updated to dynamically allocate candidate target points. The neighborhood global information sharing mechanism is designed to resolve multi-robot target conflict issues. When multiple robots compete for the same target point, the robot closer to the target point has higher priority and gains allocation rights. If a lower-priority robot has no candidate target points, a new target is randomly selected for it; otherwise, its candidate target point is selected as the new target.

[0073] In this step, considering that under the initial allocation, multiple robots may be assigned the same target point, leading to target conflicts, there will also be "unassigned target points." Within a neighborhood, when a robot A discovers an "unassigned target point," its gradient grad... A Update it to 0 and set it as its own candidate target point. If there is another robot B in the neighborhood that does not find an "unassigned target point" in its neighborhood, it will search all its neighbors, find the neighbor with the lowest gradient (let's say A), update its candidate target point to A's candidate target point, and update the gradient to grad. A +1; The gradient algorithm is designed to enable robots to quickly occupy unallocated target point resources through information sharing.

[0074] S5: Design target exchange rules based on neighborhood communication, and minimize the total time and total path distance of cluster formation through iterative exchange operations.

[0075] In one specific embodiment of the present invention, Figure 3 This is the complete operation flow of a single robot in one loop. The robot's movement requires motion control to update its position and orientation, and determines whether the conditions for global target adjustment are triggered. If the conditions are met, the nearest unoccupied target point is found, and the targets are adjusted sequentially or in reverse order, with target locking set for robots participating in the global target update. If the conditions are not met, the process proceeds directly to the next step. Next, it checks whether the target is locked. If not, the gradient is updated and the target exchange algorithm is executed; otherwise, it waits for the target to unlock.

[0076] like Figure 4 As shown, a cooperative encirclement method based on the above-mentioned cluster formation method includes the following steps:

[0077] S1: Initial allocation of the encirclement and capture groups.

[0078] After initializing the positions of the pursuers and escapees, the pursuit groups are initially assigned according to the principle of proximity. If the number of pursuers assigned is greater than the expected number of escapees, the nearest expected number of targets are selected. If there are not enough targets, they are supplemented from the pursuers who have not been assigned targets.

[0079] like Figure 5 As shown, during the initial allocation of the encirclement groups, the distance between each pursuer and each escapee is first calculated, resulting in a pursuer-escapee distance matrix. Each pursuer is then assigned the nearest escapee. For each escapee, only the nearest num_group of pursuers are retained. Here, num_group is typically the number of pursuers divided by the number of escapees, rounded down. At this point, some escapees may be assigned more pursuers than num_group, or some may be assigned fewer. In such cases, the excess pursuers are first canceled from allocation. Then, for escapees with insufficient pursuers, nearby pursuers who have not yet been assigned a target are selected as replacements until each escapee is assigned the required number of pursuers. Under this allocation, each escapee is assigned a balanced number of pursuers, and the correspondence between pursuers and escapees remains unchanged throughout the entire encirclement process.

[0080] S2: Determination of conditions for encirclement and capture.

[0081] Determine whether the encirclement conditions have been met. If so, the encirclement is suspended or a new encirclement formation is to be established; otherwise, determine whether the target location needs to be updated.

[0082] For traditional multi-target manhunt missions, each escapee is eventually surrounded by different pursuers forming a circle. The criterion for determining whether the manhunt conditions have been met is as follows: Define the set of escapees. , Let represent any escapee, and let It is to capture the escapees A collection of hunter robots. Indicates the escapee The maximum distance to all its pursuers. The maximum distance between the escapee and his pursuers can be expressed as: When the encirclement is complete: [Regarding...] , must meet , This refers to the set successful capture distance threshold. Furthermore, it is required that the capture robots are evenly distributed on the capture circle centered on the escapee, ideally forming an inscribed regular polygon, creating a closed enclosing circular region. In this embodiment, for the escapee clustering situation, the invention sets the center of the capture circle as the average position of the escapee cluster, and the capture radius is set as the maximum distance from the cluster center to each escapee within the cluster. The sum of these values, and the target location for the encirclement, is set as the vertex of the inscribed polygon on the encirclement circle formed by the cluster center and radius.

[0083] The criteria for determining whether the target location needs to be updated for encirclement are as follows:

[0084] The target position remained stable for an extended period of time.

[0085] S3: Dynamic clustering generates target points.

[0086] Based on the distance matrix of the escapee's location, a clustering algorithm is used to group the escapees, and each group corresponds to a capture circle. The capture circle is described by its capture radius and center. To prevent frequent changes in the target shape of the capture circle due to the escapee's slight movements, this invention records the capture radius and center of any cluster at the previous moment. If the recalculated center offset and radius change are both less than a set threshold, the historical capture circle parameters are used. To avoid overlapping capture circles generated by the clustering algorithm, after clustering is completed, this invention iteratively adjusts the position of the new cluster centers globally. During the iteration process, if an overlap is detected in a newly added cluster center, the new cluster center needs to be translated along the vector direction connecting the centers of conflicting clusters until all cluster circles are non-overlapping. Finally, based on the cluster centers and cluster radii of each group, a set of target points without conflict interference is generated.

[0087] like Figure 6 As shown, when calculating the escapee cluster and the target point set, the distance matrix between escapees is first calculated and the clustering is completed. The center and radius of each cluster are calculated, and it is determined whether the current cluster exists. If it does not exist, the cluster is updated. If it exists, it is further determined whether the changes in the cluster center and radius are both less than the threshold. If so, the cluster center and radius obtained in the previous round of clustering are used. Otherwise, the new cluster center and radius are used.

[0088] Based on the latest clustering, the positions of the cluster centers are iteratively adjusted globally to ensure that all cluster circles do not overlap. Then, the initial allocation of subsequent target points is performed based on the adjusted cluster centers and radii.

[0089] S4: Target assignment within clusters.

[0090] Within each cluster, target assignment from robot to target point is performed. For example... Figure 7 As shown, it is determined whether the current cluster exists in the historical cluster group. If so, the historical target point allocation scheme of the cluster is reused. Otherwise, the distance matrix between the hunter and the target point is calculated and the nearest hunting point is initially assigned to each hunter. For target points with conflict, the hunter farthest from the target point is selected and reserved for allocation, and the rest are released. The released hunters reselect the nearest target point from the remaining target points. The iterative operation continues until the target point allocation of the cluster is completed.

[0091] S5: Use the same motion control logic as the above cluster formation algorithm to update the robot's position and attitude.

[0092] Figures 8 to 10 In the diagram, the robot is displayed in blue when it has not reached the designated target point, and in green when it has reached the target point. Figure 8 The process of forming a “ZJU” formation for a robot swarm: The robots are initially distributed in the lower left corner. The target allocation is initialized using a central allocation method. After processes such as motion control and target exchange, the target formation “ZJU” is finally formed. Figure 9 The process of switching the robot swarm formation to "ROBOT": After forming "ZJU", the robot swarm completes the grouping according to the principle of "pairing up the nearest one according to the horizontal / vertical coordinates", and then randomly assigns targets within the group. Through motion control, target exchange and other methods, the formation is successfully switched to "ROBOT". Figure 10 The robot swarm is formed into a simple "eagle-shaped" polygon through processes such as "initial target allocation using a central allocation algorithm", motion control, global target adjustment, and target exchange.

[0093] Figures 11 to 14 In the diagram, the capturing robot is initially displayed in blue, the escape robot in red, the capturing robot turns green after reaching the designated position within the encirclement circle, and the escape robot turns purple after being completely captured within the encirclement circle. Figure 11 and Figure 12 The diagrams illustrate the clustering, merging, and decomposition processes that occur during the manhunt due to changes in the escapees' locations. Solid circles and hollow circles represent the manhunters and escapees, respectively. Figure 13 The process for completing the coordinated encirclement mission is as follows: At the initial moment, the escapees and the pursuers are randomly scattered. The encirclement groups are determined based on the distance matrix of the escapees / pursuers at the initial moment. The escapees are clustered into groups using a clustering algorithm to generate encirclement target points. The pursuers in the corresponding encirclement groups use the cluster formation algorithm described above to form the shape of each target (all the encirclement circles), thereby achieving the effect of coordinated encirclement. Figure 14 In the process, the movement of some escapees altered the clustering of the pursuers, and the pursuers, by dynamically adjusting the formation, successfully completed the task of recapturing all the escapees, demonstrating the flexibility of the adaptive pursuit of this invention.

[0094] This embodiment also provides a robot cooperative trapping system for implementing the above-described trapping method, comprising:

[0095] The target allocation module is used to initially allocate target points based on the target formation and by selecting an allocation strategy.

[0096] The dynamic motion control module is used to perform dynamic motion control on the robot based on the differential speed model until the target formation is achieved.

[0097] The global optimization module employs a global target redistribution mechanism to optimize the control logic in the dynamic motion control module.

[0098] The collaborative encirclement module assigns a balanced number of encirclers to each escapee and performs dynamic clustering based on the distance matrix of the escapee's location. It calculates the center and radius of each cluster and adjusts the cluster center position through global iterative optimization to avoid overlapping encirclement circles. It generates target points with an equal number of encirclers in the circumference of the encirclement circle.

[0099] The communication network module adopts a global information sharing mechanism and a target exchange mechanism within the neighborhood, and optimizes the control logic in the dynamic motion control module.

[0100] The visual monitoring interface module displays the formation and encirclement process in real time.

[0101] For the system embodiments, since they basically correspond to the method embodiments, relevant details can be found in the descriptions of the method embodiments; the implementation methods of the modules will not be repeated here. The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of the present invention according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0102] The system embodiments of the present invention can be applied to any device with data processing capabilities, such as a computer or other similar device. The system embodiments can be implemented in software, hardware, or a combination of both. Taking software implementation as an example, as a logical device, it is formed by the processor of any data processing device loading the corresponding computer program instructions from non-volatile memory into memory for execution.

[0103] It should also be noted that the target-assignment-based robot swarm formation method in the above embodiments can essentially be executed by a computer program. Therefore, similarly, based on the same inventive concept, another preferred embodiment of the present invention also provides a computer electronic device corresponding to the method provided in the above embodiments, which includes a memory and a processor;

[0104] The memory is used to store computer programs;

[0105] The processor is configured to implement the target allocation-based robot swarm formation method in the above embodiments when executing the computer program.

[0106] Furthermore, the logical instructions in the aforementioned memory can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium.

[0107] Therefore, based on the same inventive concept, another preferred embodiment of the present invention also provides a computer-readable storage medium corresponding to the method provided in the above embodiments. The storage medium stores a computer program that, when executed by a processor, can implement the target allocation-based robot swarm formation method in the above embodiments.

[0108] It is understood that the computer-readable storage medium can be an internal storage unit of any data processing device described in any of the foregoing embodiments, such as a hard disk or memory. The computer-readable storage medium can also be an external storage device of any data processing device, such as a plug-in hard disk, smart media card (SMC), SD card, flash card, etc., equipped on the device. Furthermore, the computer-readable storage medium can include both internal storage units and external storage devices of any data processing device. The computer-readable storage medium is used to store the computer program and other programs and data required by the data processing device, and can also be used to temporarily store data that has been output or will be output.

[0109] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the invention. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the invention. Therefore, all technical solutions obtained through equivalent substitution or transformation fall within the protection scope of the present invention.

Claims

1. A robot swarm formation method based on target assignment, characterized in that, include: Each robot is assigned an initial target index based on the target formation of the cluster formation, and the robot is dynamically controlled based on the differential speed model until the target formation is achieved. In the process of achieving the target formation, a global target redistribution mechanism, a neighborhood-based global information sharing mechanism, and a target exchange mechanism are introduced. The global target redistribution mechanism refers to the following: when the target formation is a closed polygon with adjacent target points distributed at equal intervals and the proportion of robots that have reached the target points exceeds the first threshold, robots that have not reached the target points and whose distance exceeds the second threshold are recorded as lagging robots. The nearest target point is selected for the lagging robots, and the selected target point is used as the starting point and the nearest unoccupied target point to the starting point is used as the ending point. The target points of the robots on the path from the starting point to the ending point are redistributed and locked in sequence to achieve target redistribution and accelerate formation convergence. The neighborhood global information sharing mechanism refers to sharing global information within the neighborhood through robot neighborhood communication, marking unassigned target points discovered by the robot as candidate target points, and updating the robot gradient value to dynamically assign candidate target points; The target exchange mechanism refers to iteratively optimizing the target points of each robot within the neighborhood, with the goal of minimizing the total formation time and total path distance.

2. The robot swarm formation method based on target allocation according to claim 1, characterized in that, Each robot is assigned an initial target index based on the target formation of the cluster, including: When the target formation is a polygonal combination with the center line parallel to the coordinate axis, it is first clustered and grouped according to the robot's coordinate position, and then the center allocation strategy within the group or the random allocation strategy is executed. The remaining target formations will be assigned to a central distribution strategy; The specific center allocation strategy is as follows: first, calculate the center position of the robot cluster, and then allocate the nearest idle robot to each target point in order of distance from the target point to the cluster center position.

3. The robot swarm formation method based on target allocation according to claim 1, characterized in that, Dynamic motion control of the robot based on the differential speed model until the target formation is achieved includes: The robot's guiding velocity component and collision avoidance velocity component are calculated and synthesized, and then weighted with the velocity of the previous moment to obtain the current combined velocity; The resultant velocity is converted into the left and right wheel velocities. A limiting constraint on the wheel velocities is introduced. The linear velocity and angular velocity of the robot are solved by the forward kinematics model, and the robot's position and attitude are updated.

4. The robot swarm formation method based on target allocation according to claim 1, characterized in that, In the global target redistribution mechanism, once the robot reaches the redistributed target point, it is unlocked. Target points that are locked do not participate in the target exchange mechanism.

5. A robot cooperative encirclement method based on the robot swarm formation method of claim 1, characterized in that, Includes the following steps: (1) Based on the distance matrix between the pursuers and the escapees, a balanced number of pursuers are assigned to each escapee. The correspondence between the pursuers and the escapees remains unchanged throughout the entire pursuit process. (2) Dynamic target point generation: Dynamic clustering and grouping are performed based on the distance matrix of the escapee's location. The center and radius of each cluster are calculated. The cluster center position is adjusted through global iterative optimization to avoid overlapping of the encirclement circles. Target points equal to the number of encirclers are generated in the circumference of the encirclement circle. (3) Initial allocation of target points: If the cluster exists in the historical group, the historical target point allocation scheme is reused; otherwise, the distance priority principle is used to allocate initial target points to the hunters within the cluster. When target points conflict, the target point is allocated to the hunter farthest from the target point, and the remaining hunters select the nearest target point from the remaining target points. (4) Motion control: The robot swarm formation method described in claim 1 is adopted, based on the global target redistribution mechanism, the global information sharing mechanism in the neighborhood and the target exchange mechanism, to control the hunters corresponding to each encirclement circle to move towards the target point to form the target formation of the encirclement circle; (5) Real-time judgment of the encirclement status. If the encirclement is completed, the process ends; otherwise, return to step (2) at each time the target location is updated, until the encirclement is completed.

6. The robot cooperative encirclement method according to claim 5, characterized in that, The process of globally iteratively optimizing and adjusting the cluster center positions includes: Detect the center and radius of the new cluster and the existing clusters, and determine whether the capture circle of the new cluster overlaps with the capture circle of the existing clusters. If so, translate the center of the new cluster along the vector direction of the line connecting the centers of the conflicting clusters until the capture circles of all clusters no longer overlap.

7. The robot cooperative encirclement method according to claim 5, characterized in that, The dynamic clustering process also includes: Determine whether the current cluster is included in the cluster grouping of the previous time step. If so, and the changes in the cluster center and radius are both less than the threshold, then use the cluster center and radius of this cluster from the previous round of cluster grouping. Otherwise, use the currently calculated cluster center and radius.

8. A robot cooperative trapping system for implementing the robot cooperative trapping method of claim 5, characterized in that the system... include: The target allocation module is used to initially allocate target points based on the target formation and by selecting an allocation strategy. The dynamic motion control module is used to perform dynamic motion control on the robot based on the differential speed model until the target formation is achieved. The global optimization module employs a global target redistribution mechanism to optimize the control logic in the dynamic motion control module. The collaborative encirclement module assigns a balanced number of encirclers to each escapee and performs dynamic clustering based on the distance matrix of the escapee's location. It calculates the center and radius of each cluster and adjusts the cluster center position through global iterative optimization to avoid overlapping encirclement circles. It generates target points with an equal number of encirclers in the circumference of the encirclement circle. The communication network module adopts a global information sharing mechanism and a target exchange mechanism within the neighborhood, and optimizes the control logic in the dynamic motion control module. The visual monitoring interface module displays the formation and encirclement process in real time.

9. A computer-readable storage medium, characterized in that, The storage medium stores a computer program that, when executed by a processor, implements the target-assignment-based robot swarm formation method as described in any one of claims 1 to 4.

10. A computer electronic device, characterized in that, Including memory and processor; The memory is used to store computer programs; The processor is configured to, when executing the computer program, implement the target-assignment-based robot swarm formation method as described in any one of claims 1 to 4.

Images