A robot swarm dynamic alliance hunting mission planning method based on CBCA algorithm

Abstract

The present invention provides a robot cluster dynamic alliance capture task planning method based on a consensus coalition algorithm (CBCA), belonging to the technical field of multi-robot collaborative task planning. The present invention adopts the CBCA algorithm to perform task planning on the robot cluster, and realizes the capture task allocation of multiple mobile robots to multiple dynamic targets. The capture task planning is divided into two stages: task pre-allocation and capture process replanning. In the pre-allocation stage, robots form temporary groups based on limited perception range and communication range, construct their own task packages, and resolve conflicts through the consistency principle to form a capture alliance. The robot loyalty is set to ensure the convergence of the capture task replanning algorithm, thereby realizing the orderly and efficient capture of multiple dynamic targets by the robot cluster, and improving the average execution efficiency and average performance of the robots.

Description

Technical Field

[0001] The present invention relates to the technical field of multi-robot collaborative task planning, and in particular to a robot cluster dynamic alliance roundup task planning method based on a CBCA algorithm. Background Art

[0002] Robot collaborative capture mission planning algorithms have significant application value in target tracking and interception, search and rescue, and competitive competition, and have long been a research hotspot. Currently, most research on multi-robot capture mission planning relies on centralized allocation algorithms, where a global coordinator performs unified planning. However, if new tasks or new robots enter the environment, the difficulty of solving centralized algorithms increases dramatically, impacting the scalability and efficiency of the allocation algorithms. Therefore, centralized algorithms are not suitable for dynamically changing capture mission allocation scenarios.

[0003] In recent years, the consensus coalition algorithm (CBCA) in task planning, which eliminates the need for a centralized coordinator and optimizes solutions solely through information transfer and negotiation between adjacent participants, has been widely studied in robotic swarm task planning. This algorithm can efficiently enable robots to obtain target assignments and form coalitions to execute their respective tasks. However, the consensus coalition algorithm is currently only applicable to scenarios where static task objectives are assigned and requires advance knowledge of all task objectives. This makes it less applicable to scenarios where robots have limited perception capabilities and requires improvement. Summary of the Invention

[0004] To address the aforementioned technical issues, a method for planning dynamic coalition capture tasks for robot swarms based on the CBCA algorithm is provided. This method designs a consistent coalition pre-allocation algorithm for robots with limited perception and communication capabilities, as well as a dynamic replanning algorithm during the capture process. A loyalty model is proposed to ensure algorithm convergence, enabling the orderly and efficient capture of multiple dynamic targets by robot swarms, thereby improving the robots' average execution efficiency and performance.

[0005] The technical means adopted in the present invention are as follows:

[0006] A robot swarm dynamic alliance roundup mission planning method based on the CBCA algorithm includes:

[0007] S1. Set the robot capture and mission objectives;

[0008] S2. Setting constraints for the capture robot;

[0009] S3, construct the robot kinematic model;

[0010] S4. Use artificial potential field method to navigate the robot's movement. The combined force of gravity and repulsion guides the robot from the starting point to the end point while avoiding other robots in the movement trajectory.

[0011] S5. Construct the objective function;

[0012] S6. Setting CBCA algorithm parameters;

[0013] S7, build a task package based on the local detection information and limited communication range of the robot, and form an alliance to carry out collaborative roundup mission planning;

[0014] S8, conflict resolution based on local information interaction;

[0015] S9. Task re-planning.

[0016] Furthermore, the step S1 specifically includes:

[0017] S11, set the capture robot to Indicates n r A collection of capture robots, each robot is represented by the three-element method:

[0018] X i ,E i ',l i >(1)

[0019] Among them, X i is the position coordinate of the robot, E i 'A list of resources carried by the robot, l i is the maximum number of tasks that robot i can accept;

[0020] S12. Set the task goal as Indicates n e There are dynamically moving task targets, and each task is represented by the four-element method:

[0021] X j ,E j ,L j ,V a >(2)

[0022] Among them, X j is the location of the mission target, E j Indicates the list of resources required for the task, L j is the number of robots required for task j, indicating that task j requires multiple robots to complete the roundup task. is the initial value set of the task goal;

[0023] S13. Express the various resource requirements of each task objective as a vector:

[0024]

[0025] in, represents the number of p-th resource required to capture mission target j;

[0026] S14. In the capture mission, the different types and quantities of resources carried by each capture robot are represented by the following z-dimensional vector:

[0027]

[0028] in, Indicates the number of the p-th resource it carries.

[0029] Furthermore, the step S2 specifically includes:

[0030] S21. Set the robot's mileage constraints. The maximum mileage that each robot can move is limited, that is:

[0031]

[0032] in, Indicates the maximum mileage the robot can move;

[0033] S22. Set resource constraints for the robots. Each capture robot carries limited resources that are consumed as they are used. Considering that each robot must execute multiple assigned tasks in sequence, its resources must meet the following requirements:

[0034] E(p i )≤E i '(6)

[0035] Among them, E(p i ) indicates that robot i executes its task sequence p i The resources consumed by all the above targets;

[0036] S23. Set the resource constraints of the robot alliance. The total amount of resources carried by the robot alliance built for the discovered target j should not be less than the amount of resources required by the target, that is:

[0037]

[0038] Among them, Φ j Represents the robot alliance formed for the capture task j.

[0039] Furthermore, the robot kinematic model constructed in step S3 is specifically:

[0040]

[0041]

[0042]

[0043] in, represents the position coordinates of the i-th robot at the k-th step, represents the steering angle of the i-th robot, represents the angular velocity of the i-th robot, and Δt represents the control time step.

[0044] Furthermore, the step S4 specifically includes:

[0045] S41. Express gravity as:

[0046]

[0047] Among them, k a represents the scale factor of the gravitational potential field, X i =(x i ,y i ) represents the position coordinates of the i-th robot, and the target point coordinates are

[0048] S42. Express the repulsive force as:

[0049]

[0050]

[0051] Among them, k r is the repulsion coefficient, d o Represents the influence radius of the repulsive field, d(X i ,X q ) is the distance between robot i and the qth robot, X q =(x q ,y q ) is the coordinate of the current mobile robot; the factor is introduced into the repulsion function The repulsive force is composed of two parts, but they point in different directions. is the force from robot i to the target, and is the force directed from other mobile robots toward robot i;

[0052] S43. When the capture robot reaches the target point, it needs to form a capture formation around the target. Define X m =(x m ,y m) is the coordinate of the mth capture point around the target; within the capture range, the gravitational field of the original target point needs to be changed to the gravitational field of the capture point, and the attraction model of the capture point is obtained:

[0053]

[0054] Furthermore, the step S5 specifically includes:

[0055] S51, the profit r of robot i performing task target j ij (p i ) decays with time:

[0056]

[0057] Among them, V aj The initial value of the jth goal, is the sequence of all tasks that robot i needs to complete in chronological order, t o The start time assigned to the task, For the robot to follow the task path p i Execution target T j Estimated time; 0≤β j ≤1 is the task time discount factor, which reflects how quickly the goal value decreases over time;

[0058] S52, robot i along path p i The performance is:

[0059]

[0060] Among them, ξ ij =d ij γ is the robot’s travel cost, d ij is the straight-line distance between robot i and task j, and γ is the distance cost coefficient;

[0061] S53. Combined with the task allocation requirements, the objective function is described as:

[0062]

[0063] st(5)(6)(7)

[0064]

[0065]

[0066]

[0067] Among them, x ij ∈{1,0} is the decision variable in task allocation, xij =1 indicates that the robot performs task j, x ij =0 means the robot does not perform this task.

[0068] Furthermore, the step S6 specifically includes:

[0069] S61, set the task package to is the set of tasks assigned to robot i;

[0070] S62. Set the winning robot matrix n r ×n e dimensional matrix, where indicates that robot i believes that robot k is the winner of the bid for task j, otherwise Z i The sum of the number of non-zero elements in the jth column represents the total number of robots that robot i believes are performing task j;

[0071] S63. Set the winning bid value matrix to Store the winning bid value for task j from the perspective of robot i Each element in the matrix and Z i One-to-one correspondence;

[0072] S64. Set the time stamp list Record the time when robot i obtains updated information from other adjacent robots, where s ik represents the time when robot i obtains the latest information from robot k.

[0073] Furthermore, the step S7 specifically includes:

[0074] S71. Every time robot i selects a task, it selects task j with the largest incremental benefit and adds it to task list b. i The specific rules are as follows: each robot joins b according to the new task j' i The marginal gain of j' can be obtained by comparing the benefits before and after:

[0075]

[0076] Among them, n represents j' in p i All possible positions in |p i | represents the length of the path list; Indicates inserting j' into the path list p i The nth position in the original path list and the elements after it keep their original order and move backward in sequence;

[0077] S72. Each target in the capture mission must be completed by three or more robots. If the number of robots bidding to perform task j meets the number of alliance robots, that is, At this point, the number of robots assigned to perform task j has met the requirement, and the robot compares its own peripheral gain with the minimum winning bid value of the current task; if Then robot i can bid and replace the minimum winning bid value with its own bid value; otherwise, agent i gives up task j; when the bid value of robot i is the same as the minimum winning bid value, the robot with the smaller identity is selected; when the task package is full or there is no task to be selected, the task package construction process ends.

[0078] Furthermore, the step S8 specifically includes:

[0079] S81. In the conflict resolution phase, if the winning bidder identifiers from the perspectives of the robot and its neighboring robots are the same but with different timestamps, each robot updates its own winning bidder matrix, thereby reaching a consensus with its neighbors and confirming the final winning bidder.

[0080] S82. If the winning bidder identifications from the perspectives of the robot and the adjacent robot are different, determine whether the number of winning bidders for the task bid by the conflicting robot has reached the upper limit. If not, consider the identities of all winning bidders valid.

[0081] S83. If the number of conflicting bidders exceeds the upper limit of the number required for the task, each robot removes the other bidders with the lowest efficiency from its perspective and updates its own bidder matrix. If the robot itself is the bidder with the lowest efficiency, it loses its bidder status and needs to reset the position of this task in its task package sequence, returning to the task package construction phase.

[0082] S84. Each group uses the above two steps to obtain a pre-allocation result within the group based on the currently known task information, forming multiple alliances, represented as

[0083] Furthermore, the step S9 specifically includes:

[0084] S91. Establish the robot loyalty function model as follows:

[0085]

[0086] in, It represents the percentage of various resources owned by robot i in the total resources of the alliance, reflecting the importance of robot i to the alliance; d ic =t ic γ is the mileage consumed by robot i in performing the current task, t icis the time the robot has traveled; μ i =S(b i ) is the default cost of robot i choosing to change to other task targets, S(b i ) is the set of tasks b that robot i needs to perform i This item means that if robot i chooses to breach the contract, it will have to bear the loss of income from all tasks it has not performed;

[0087] S92. Assume that a robot wants to change the execution order of the current task, which needs to satisfy the following formula:

[0088] S i (p i ')>S i (p i )+σ i (twenty three)

[0089] That is, the task path can be changed only when the new task can bring enough efficiency to the robot, thus avoiding frequent task switching and ensuring the convergence of the algorithm;

[0090] S93. Based on the loyalty model, task replanning is specifically divided into the following situations:

[0091] Case 1: If an alliance discovers a new target that can be captured and no other alliance is currently capturing this target, the alliance will decide whether to re-plan based on its current capture status: if the current target is already under siege, the original mission path will be maintained; otherwise, the new target's value will be determined to be greater than the current target's value, and the target with greater value will be selected for capture;

[0092] Case 2: If the alliances that were originally disconnected enter each other's communication range during their respective roundups, there are two cases to discuss:

[0093] If an alliance has already besieged the current mission target, it will not participate in mission replanning;

[0094] If each alliance fails to besiege the target, they will merge into a new alliance, re-establish the task package, and resolve the consistency conflict to obtain a new task path p i ' and winning bid income S i (p i '), and judge whether to execute the roundup mission according to the new mission path according to formula (23).

[0095] Compared with the prior art, the present invention has the following advantages:

[0096] 1. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm provided by the present invention improves the decentralized consistency alliance algorithm, and can construct task packages based on the local detection information and limited communication range of the robots to form an alliance for collaborative roundup task planning.

[0097] 2. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm provided by the present invention divides the roundup task planning into two stages: global pre-allocation and roundup process replanning, which improves the adaptability of the allocation algorithm to dynamically changing scenarios.

[0098] 3. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm provided by the present invention proposes a mathematical model of the robot's loyalty to the task target, ensuring the convergence of the re-planning algorithm.

[0099] Based on the above reasons, the present invention can be widely promoted in fields such as multi-robot collaborative task planning. BRIEF DESCRIPTION OF THE DRAWINGS

[0100] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0101] Figure 1 Flow chart of the method of the present invention.

[0102] Figure 2 A state diagram of the capture process provided by an embodiment of the present invention.

[0103] Figure 3 The number of task objectives n provided by the embodiment of the present invention e =30 when the average system performance and execution time curve. DETAILED DESCRIPTION

[0104] It should be noted that, in the absence of conflict, the embodiments and features of the embodiments of the present invention can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0105] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. The following description of at least one exemplary embodiment is actually only illustrative and is in no way intended to limit the present invention and its application or use. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0106] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.

[0107] Unless otherwise specifically stated, the relative arrangement of the parts and steps, numerical expressions and numerical values ​​described in these embodiments do not limit the scope of the present invention. At the same time, it should be clear that, for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship. The techniques, methods and equipment known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and equipment should be considered as part of the authorization specification. In all examples shown and discussed here, any specific value should be interpreted as merely exemplary and not as a limitation. Therefore, other examples of the exemplary embodiments may have different values. It should be noted that similar numbers and letters represent similar items in the following figures, so once an item is defined in one figure, it does not need to be further discussed in subsequent figures.

[0108] In the description of the present invention, it should be understood that the directions or positional relationships indicated by directional words such as "front, back, up, down, left, right", "horizontal, vertical, vertical, horizontal" and "top, bottom" are usually based on the directions or positional relationships shown in the accompanying drawings. They are only for the convenience of describing the present invention and simplifying the description. Unless otherwise specified, these directional words do not indicate or imply that the device or element referred to must have a specific direction or be constructed and operated in a specific direction. Therefore, they cannot be understood as limiting the scope of protection of the present invention: the directional words "inside and outside" refer to the inside and outside relative to the outline of each component itself.

[0109] For ease of description, spatially relative terms such as "above", "above", "on the upper surface of", "above", etc. may be used herein to describe the spatial positional relationship of a device or feature to other devices or features as shown in the figures. It should be understood that spatially relative terms are intended to include different orientations of the device in use or operation in addition to the orientation described in the figures. For example, if the device in the drawings is inverted, the device described as "above other devices or structures" or "above other devices or structures" will be positioned as "below other devices or structures" or "below their position devices or structures". Thus, the exemplary term "above" can include both "above" and "below". The device can also be positioned in other different ways (rotated 90 degrees or in other orientations), and the spatially relative descriptions used here are interpreted accordingly.

[0110] In addition, it should be noted that the use of terms such as "first" and "second" to limit components is only for the convenience of distinguishing the corresponding components. Unless otherwise stated, the above terms have no special meaning and therefore cannot be understood as limiting the scope of protection of the present invention.

[0111] like Figure 1 As shown, the present invention provides a robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm, comprising:

[0112] S1. Set the robot capture and mission objectives;

[0113] S2. Setting constraints for the capture robot;

[0114] S3, construct the robot kinematic model;

[0115] S4. Use artificial potential field method to navigate the robot's movement. The combined force of gravity and repulsion guides the robot from the starting point to the end point while avoiding other robots in the movement trajectory.

[0116] S5. Construct the objective function;

[0117] S6. Setting CBCA algorithm parameters;

[0118] S7, build a task package based on the local detection information and limited communication range of the robot, and form an alliance to carry out collaborative roundup mission planning;

[0119] S8, conflict resolution based on local information interaction;

[0120] S9. Task re-planning.

[0121] In specific implementation, as a preferred embodiment of the present invention, step S1 specifically includes:

[0122] S11, set the capture robot to Indicates n r A collection of capture robots, each robot is represented by the three-element method:

[0123] X i ,E i ',l i >(1)

[0124] Among them, X i is the position coordinate of the robot, E i 'A list of resources carried by the robot, l i is the maximum number of tasks that robot i can accept;

[0125] S12. Set the task goal as Indicates n e There are dynamically moving task targets, and each task is represented by the four-element method:

[0126] X j ,E j ,L j ,V a >(2)

[0127] Among them, X j is the location of the mission target, E j Indicates the list of resources required for the task, L j is the number of robots required for task j, indicating that task j requires multiple robots to complete the roundup task. is the initial value set of the task goal;

[0128] S13. Express the various resource requirements of each task objective as a vector:

[0129]

[0130] in, represents the number of p-th resource required to capture mission target j;

[0131] S14. In the capture mission, the different types and quantities of resources carried by each capture robot are represented by the following z-dimensional vector:

[0132]

[0133] in, Indicates the number of the p-th resource it carries.

[0134] In specific implementation, as a preferred embodiment of the present invention, step S2 specifically includes:

[0135] S21. Set the robot's mileage constraints. The maximum mileage that each robot can move is limited, that is:

[0136]

[0137] in, Indicates the maximum mileage the robot can move;

[0138] S22. Set resource constraints for the robots. Each capture robot carries limited resources that are consumed as they are used. Considering that each robot must execute multiple assigned tasks in sequence, its resources must meet the following requirements:

[0139] E(p i )≤E i '(6)

[0140] Among them, E(p i ) indicates that robot i executes its task sequence p i The resources consumed by all the above targets;

[0141] S23. Set the resource constraints of the robot alliance. The total amount of resources carried by the robot alliance built for the discovered target j should not be less than the amount of resources required by the target, that is:

[0142]

[0143] Among them, Φ j Represents the robot alliance formed for the capture task j.

[0144] In specific implementation, as a preferred embodiment of the present invention, the robot kinematic model constructed in step S3 is specifically:

[0145]

[0146]

[0147]

[0148] in, represents the position coordinates of the i-th robot at the k-th step, represents the steering angle of the i-th robot, represents the angular velocity of the i-th robot, and Δt represents the control time step.

[0149] In specific implementation, as a preferred embodiment of the present invention, step S4 specifically includes:

[0150] S41. Express gravity as:

[0151]

[0152] Among them, k a represents the scale factor of the gravitational potential field, X i =(x i ,y i ) represents the position coordinates of the i-th robot, and the target point coordinates are

[0153] S42. Express the repulsive force as:

[0154]

[0155]

[0156] Among them, k r is the repulsion coefficient, d o Represents the influence radius of the repulsive field, d(X i ,X q ) is the distance between robot i and the qth robot, X q =(x q ,y q ) is the coordinate of the current mobile robot; the factor is introduced into the repulsion function The repulsive force is composed of two parts, but they point in different directions. is the force from robot i to the target, and is the force directed from other mobile robots toward robot i;

[0157] S43. When the capture robot reaches the target point, it needs to form a capture formation around the target. Define X m =(x m ,y m ) is the coordinate of the mth capture point around the target; within the capture range, the gravitational field of the original target point needs to be changed to the gravitational field of the capture point, and the attraction model of the capture point is obtained:

[0158]

[0159] In specific implementation, as a preferred embodiment of the present invention, step S5 specifically includes:

[0160] S51, the profit r of robot i performing task target j ij (p i ) decays with time:

[0161]

[0162] Among them, V ajThe initial value of the jth goal, is the sequence of all tasks that robot i needs to complete in chronological order, t o The start time assigned to the task, For the robot to follow the task path p i Execution target T j Estimated time; 0≤β j ≤1 is the task time discount factor, which reflects how quickly the goal value decreases over time;

[0163] S52, robot i along path p i The performance is:

[0164]

[0165] Among them, ξ ij =d ij γ is the robot’s travel cost, d ij is the straight-line distance between robot i and task j, and γ is the distance cost coefficient;

[0166] S53. Combined with the task allocation requirements, the objective function is described as:

[0167]

[0168] st(5)(6)(7)

[0169]

[0170]

[0171]

[0172] Among them, x ij ∈{1,0} is the decision variable in task allocation, x ij =1 indicates that the robot performs task j, x ij =0 means the robot does not perform this task.

[0173] In specific implementation, as a preferred embodiment of the present invention, step S6 specifically includes:

[0174] S61, set the task package to is the set of tasks assigned to robot i;

[0175] S62. Set the winning robot matrix n r ×n e dimensional matrix, where indicates that robot i believes that robot k is the winner of the bid for task j, otherwise Zi The sum of the number of non-zero elements in the jth column represents the total number of robots that robot i believes are performing task j;

[0176] S63. Set the winning bid value matrix to Store the winning bid value for task j from the perspective of robot i Each element in the matrix and Z i One-to-one correspondence;

[0177] S64. Set the time stamp list Record the time when robot i obtains updated information from other adjacent robots, where s ik represents the time when robot i obtains the latest information from robot k.

[0178] The multi-target capture task assignment method based on the CBCA algorithm includes steps such as task package construction and conflict resolution. Since the detection capability of each capture robot is limited, it is necessary to pass task attribute information and bidding information through neighbors. Therefore, before task assignment, each capture robot forms multiple temporary groups based on the detection range, represented as Ψ s , including all robots that can form a communication link. The robots in the group can share task information and build their own task packages based on the shared information. The specific process of the algorithm is as follows Figure 1 shown.

[0179] In specific implementation, as a preferred embodiment of the present invention, step S7 specifically includes:

[0180] S71. Every time robot i selects a task, it selects task j with the largest incremental benefit and adds it to task list b. i The specific rules are as follows: each robot joins b according to the new task j' i The marginal gain of j' can be obtained by comparing the benefits before and after:

[0181]

[0182] Among them, n represents j' in p i All possible positions in |p i | represents the length of the path list; Indicates inserting j' into the path list p i The nth position in the original path list and the elements after it keep their original order and move backward in sequence;

[0183] S72. Each target in the capture mission must be completed by three or more robots. If the number of robots bidding to perform task j meets the number of alliance robots, that is, At this point, the number of robots assigned to perform task j has met the requirement, and the robot compares its own peripheral gain with the minimum winning bid value of the current task; if Then robot i can bid and replace the minimum winning bid value with its own bid value; otherwise, agent i gives up task j; when the bid value of robot i is the same as the minimum winning bid value, the robot with the smaller identity is selected; when the task package is full or there is no task to be selected, the task package construction process ends.

[0184] In specific implementation, as a preferred embodiment of the present invention, step S8 specifically includes:

[0185] S81. In the conflict resolution phase, if the winning bidder identifiers from the perspectives of the robot and its neighboring robots are the same but with different timestamps, each robot updates its own winning bidder matrix, thereby reaching a consensus with its neighbors and confirming the final winning bidder.

[0186] S82. If the winning bidder identifications from the perspectives of the robot and the adjacent robot are different, determine whether the number of winning bidders for the task bid by the conflicting robot has reached the upper limit. If not, consider the identities of all winning bidders valid.

[0187] S83. If the number of conflicting bidders exceeds the upper limit of the number required for the task, each robot removes the other bidders with the lowest efficiency from its perspective and updates its own bidder matrix. If the robot itself is the bidder with the lowest efficiency, it loses its bidder status and needs to reset the position of this task in its task package sequence, returning to the task package construction phase.

[0188] S84. Each group uses the above two steps to obtain a pre-allocation result within the group based on the currently known task information, forming multiple alliances, represented as

[0189] In specific implementation, as a preferred embodiment of the present invention, step S9 specifically includes:

[0190] S91. Establish the robot loyalty function model as follows:

[0191]

[0192] in, It represents the percentage of various resources owned by robot i in the total resources of the alliance, reflecting the importance of robot i to the alliance; d ic =t ic γ is the mileage consumed by robot i in performing the current task, t ic is the time the robot has traveled; μ i =S(b i) is the default cost of robot i choosing to change to other task targets, S(b i ) is the set of tasks b that robot i needs to perform i This item indicates that if robot i chooses to breach the contract, it will have to bear the loss of income from all tasks it has not performed;

[0193] S92. Assume that a robot wants to change the execution order of the current task, which needs to satisfy the following formula:

[0194] S i (p i ')>S i (p i )+σ i (twenty three)

[0195] That is, the task path can be changed only when the new task can bring enough efficiency to the robot, thus avoiding frequent task switching and ensuring the convergence of the algorithm;

[0196] S93. Based on the loyalty model, task replanning is specifically divided into the following situations:

[0197] Case 1: If an alliance discovers a new target that can be captured and no other alliance is currently capturing this target, the alliance will decide whether to re-plan based on its current capture status: if the current target is already under siege, the original mission path will be maintained; otherwise, the new target's value will be determined to be greater than the current target's value, and the target with greater value will be selected for capture;

[0198] Case 2: If the alliances that were originally disconnected enter each other's communication range during their respective roundups, there are two cases to discuss:

[0199] If an alliance has already besieged the current mission target, it will not participate in mission replanning;

[0200] If each alliance fails to besiege the target, they will merge into a new alliance, re-establish the task package, and resolve the consistency conflict to obtain a new task path p i ' and winning bid income S i (p i '), and judge whether to execute the roundup mission according to the new mission path according to formula (23).

[0201] Example 1

[0202] In this paper, the proposed algorithm is simulated and verified using MATLAB 2020b simulation software, and the performance is compared with the existing algorithm to verify the advantages of the proposed algorithm. The simulation experiment is carried out in an environment area of ​​10km×10km. The detection range radius of each robot is 1.2km, the communication radius between robots is 1.5km, and the maximum number of tasks that the robot can perform is l. i There are 3 of each.

[0203] In this embodiment, the captured target remains silent before being chased. After the capturing robots start capturing, the target begins to escape, and the escape direction is the midline direction of the maximum angle formed by the positions of adjacent capturing robots and the target position.

[0204] Table 1 Initial parameters of mission objectives

[0205]

[0206] Table 2 Robot initial parameters

[0207]

[0208] In this Example 1, there are 5 targets and 15 robots in the capture mission environment. The other parameters of the mission targets and robots are shown in Tables 1 and 2. Under the condition that the observation capability and communication range of the robots are limited, the improved CBCA algorithm is applied to a capture example scenario to illustrate the whole process of multi-robot collaborative capture. The initial positions of the robots and the capture targets are as follows: Figure 2 As shown in (a), the blue triangles represent robots and the red dots represent targets. All the capture robots in the task space are formed into two temporary groups Ψ1 and Ψ2 according to the detection range and communication distance. Figure 2 (b) is the result of the allocation of each temporary group using the CBCA algorithm, and the formed alliance is Φ1={R1,R 15 ,R5}, Φ2={R2,R4,R6}, Φ3={R7,R8,R9}, Φ4={R 13 ,R 12 ,R 11 ,R 10}. Figure 2 In (c), the encircling robots have already started encircling, and the members of the two temporary teams have entered their respective communication ranges, thus forming a new large alliance. Through the re-planning strategy of the CBCA algorithm, the new alliance has carried out another mission planning. Figure 2 (d) It can be seen that R9 joins the execution of task T5 by replanning the task path. 15 Joined the execution of task T3, R9 and R 15If the efficiency gained from executing the new goal is {204.39, 104.53}, both are greater than the efficiency gained from the original goal {110.56, 36.32}. In the alliance of T1, since the original R 15 Other tasks have been executed, and there is no capture alliance that meets the resource requirements of T1 at the time of replanning, so the T1 task is put on hold waiting for other robot alliances to discover and replan the task. Figure 2 (e) Each capture robot has reached its respective capture point and has trapped each mission target.

[0209] Example 2

[0210] In Example 2, when the number of robots in the environment is n r When the value of is changed, the average system performance and average task completion time of the algorithm proposed in the present invention are compared with those of an improved contract network algorithm. Figure 3 (a) As can be seen, the average performance of the proposed algorithm is higher than that of the contract network algorithm when the number of robots changes. Furthermore, as the number of robots increases, the proposed algorithm consistently outperforms the contract network algorithm. As the number of robots continues to increase, the curves tend to stabilize. This is because, with a sufficiently large number of robots, the task allocation algorithm gradually finds the optimal robot combination. Figure 3 (b) shows a comparison of the average robot completion times for the two algorithms. It can be seen that the average completion times for both algorithms are decreasing. This is because the increasing number of robots in the environment brings the tasks they perform closer to the robots themselves, resulting in a gradual decrease in the completion time of the robots performing the roundup task. Furthermore, the average task completion time of the proposed algorithm is lower than that of the contract network algorithm. This is because the proposed algorithm reduces the frequency of interactions between robots, resulting in a slight decrease in the solution runtime, thereby improving roundup efficiency.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the above embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A robot cluster dynamic alliance roundup task planning method based on CBCA algorithm, characterized by: include: S1. Set the robot capture and mission objectives; S2. Setting constraints for the capture robot; S3, construct the robot kinematic model; S4. Use artificial potential field method to navigate the robot's movement. The combined force of gravity and repulsion guides the robot from the starting point to the end point while avoiding other robots in the movement trajectory. S5. Construct the objective function; S6. Setting CBCA algorithm parameters; S7, build a task package based on the local detection information and limited communication range of the robot, and form an alliance to carry out collaborative roundup mission planning; S8. Conflict resolution based on local information interaction, specifically including: S81. In the conflict resolution phase, if the winning bidder identifiers from the perspectives of the robot and its neighboring robots are the same but with different timestamps, each robot updates its own winning bidder matrix, thereby reaching a consensus with its neighbors and confirming the final winning bidder. S82. If the winning bidder identifications from the perspectives of the robot and the adjacent robot are different, determine whether the number of winning bidders for the task bid by the conflicting robot has reached the upper limit. If not, consider the identities of all winning bidders valid. S83. If the number of conflicting bidders exceeds the upper limit of the number required for the task, each robot removes the other bidders with the lowest efficiency from its perspective and updates its own bidder matrix. If the robot itself is the bidder with the lowest efficiency, it loses its bidder status and needs to reset the position of this task in its task package sequence, returning to the task package construction phase. S84. Each group uses the above three steps to obtain a pre-allocation result within the group based on the currently known task information, forming multiple alliances, represented as , ; S9. Task re-planning, including: S91. Establish the robot loyalty function model as follows: （22） in, Represents a robot The percentage of each resource owned by the robot to the total resources of the alliance reflects the Importance to the Alliance; It's a robot The mileage consumed in executing the current task, is the time the robot has traveled, is the distance cost coefficient; It's a robot The default cost of choosing to change other mission targets, It's a robot The set of tasks to be performed The income of this item means that if the robot If it chooses to breach the contract, it will be required to bear the loss of income from all tasks it has not performed; S92. Assume that a robot wants to change the execution order of the current task, which needs to satisfy the following formula: （23） That is, the task path can be changed only when the new task can bring enough efficiency to the robot, thus avoiding frequent task switching and ensuring the convergence of the algorithm; For robots Along the path effectiveness; S93. Based on the loyalty function model, task replanning is specifically divided into the following cases: Case 1: If an alliance discovers a new target that can be captured and no other alliance is currently capturing this target, the alliance will decide whether to re-plan based on its current capture status: if the current target is already under siege, the original mission path will be maintained; otherwise, the new target's value will be determined to be greater than the current target's value, and the target with greater value will be selected for capture; Case 2: If the alliances that were originally disconnected enter each other's communication range during their respective roundups, there are two cases to discuss: If an alliance has already besieged the current mission target, it will not participate in mission replanning; If none of the alliances have besieged the target, they will merge into a new alliance, re-establish the mission package, and resolve the consistency conflict to obtain a new mission path. and winning bid income , according to formula (23), determine whether to execute the roundup mission according to the new mission path.

2. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 1 is characterized in that: The step S1 specifically includes: S11, set the capture robot to ,express A collection of capture robots, each robot is represented by the three-element method: （1） in, is the position coordinate of the robot, A list of resources carried by the robot. For robots The maximum number of tasks that can be accepted; S12. Set the task goal as ,express There are dynamically moving task targets, and each task is represented by the four-element method: （2） in, is the location of the mission target, Indicates the list of resources required for the task. For the task The number of robots required to represent the task Multiple robots are needed to complete the roundup mission. is the initial value set of the task goal; S13. Express the various resource requirements of each task objective as a vector: ， （3） in, Indicates the mission goal The first the number of resources; S14. In the capture mission, the different types and quantities of resources carried by each capture robot are as follows: dimensional vector to represent: ， （4） in, Indicates the number of The number of resources.

3. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 2 is characterized in that: The step S2 specifically includes: S21. Set the robot's mileage constraints. The maximum mileage that each robot can move is limited, that is: ， （5） in, Indicates the maximum mileage the robot can move; S22. Set resource constraints for the robots. Each capture robot carries limited resources that are consumed as they are used. Considering that each robot must execute multiple assigned tasks in sequence, its resources must meet the following requirements: （6） in, Represents a robot Execute its task sequence The resources consumed by all the above targets; S23. Set the robot's alliance resource constraints for the discovered target The total amount of resources carried by the constructed robot alliance should not be less than the amount of resources required by the target, that is: （7） in, Indicates the capture mission The robot alliance formed.

4. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 3 is characterized in that: The robot kinematic model constructed in step S3 is specifically: （8） （9） （10） in, Indicates the The robot in Position coordinates at step time, Indicates the The steering angle of the robot, Indicates the The angular velocity of the robot, Represents the control time step.

5. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 4 is characterized in that: The step S4 specifically includes: S41. Express gravity as: （11） in, represents the scale factor of the gravitational potential field, Indicates the The position coordinates of the robot and the target point coordinates are ; S42. Express the repulsive force as: （12） （13） in, is the repulsion coefficient, represents the influence radius of the repulsive field, It's a robot With other The distance between the robots, is the coordinate of the current mobile robot; the factor is introduced into the repulsion function , so that the capture robot will not be unable to reach the target due to excessive repulsion when it is heading to the destination; the repulsion is composed of two parts, but the directions are different. From the robot i The force directed toward the target, Is pointing from other mobile robots to the robot i force; S43. When the capture robot reaches the target point, it needs to form a capture formation around the target. For the target m The coordinates of the capture point; within the capture range, the gravitational field of the original target point needs to be changed to the gravitational field of the capture point, and the attraction model of the capture point is obtained: （14）。 6. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 5 is characterized in that: The step S5 specifically includes: S51, Robot i Execution mission objectives j Profit Decay over time: （15） in, No. The initial value of a target, For robots A sequence of all tasks that need to be completed in chronological order. The start time assigned to the task, For the robot to follow the task path Execution Goal the estimated time of is the task time discount factor, which reflects how quickly the target value decreases over time; S52, Robot Along the path The performance is: （16） in, is the robot's journey cost, For robots With the task The straight-line distance between is the distance cost coefficient; S53. Combined with the task allocation requirements, the objective function is described as: （17） st (5) (6) (7) （18） （19） （20） in, is the decision variable in task assignment, Indicates that the robot performs the task , Indicates that the robot does not perform the task.

7. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 6 is characterized in that: The step S6 specifically includes: S61, set the task package to , is assigned to the robot A set of tasks; S62. Set the winning robot matrix for dimensional matrix, where Represents a robot Think Robot It's a task The winner of the bid, otherwise ; Middle The sum of the number of nonzero elements in the column Represents a robot Think of performing tasks Total number of robots; S63. Set the winning bid value matrix to , storage robot Tasks from a perspective Winning bid value , each element in the matrix is One-to-one correspondence; S64. Set the time stamp list Recording Robot The time when updated information is obtained from other adjacent robots, Represents a robot From the robot Get the latest information at the moment.

8. The robot cluster dynamic alliance roundup task planning method based on the CBCA algorithm according to claim 7 is characterized in that: The step S7 specifically includes: S71, Robot Every time you choose a task, choose the one with the largest incremental benefit. Add to task list The specific rules are as follows: each robot will perform a new task based on join in The benefits before and after can be obtained The marginal gain is: （21） in, express exist All possible positions in Expressed as the path list length; Indicates that Insert into path list Middle The first position in the original path list The elements after the first bit remain in the original order and are shifted back in sequence; S72. Each target in the roundup mission must be completed by three or more robots in collaboration. The number of robots has met the number of alliance robots required, that is, , at this time, the execution task is assigned The number of robots has met the requirement, and the robot compares its own peripheral gain with the minimum winning bid value of the current task; if , then the robot i Can bid and replace the minimum winning bid value with its own bid value; otherwise, the robot i Abandon the mission When robots i When the bid value is the same as the minimum winning bid value, the robot with the smaller identity is selected; when the task package is full or there are no tasks to be selected, the task package construction process ends.

Images