Unmanned equipment cluster cooperative scheduling method and system based on behavior tree

Abstract

This application discloses a method and system for collaborative scheduling of unmanned equipment clusters based on behavior trees, relating to the field of equipment scheduling technology. The method includes: collecting equipment categories and basic state parameters of unmanned equipment; randomly selecting one unmanned equipment as the first unmanned equipment and obtaining an initial behavior tree; updating the weights of the root node based on the equipment category and task status of the first unmanned equipment; updating the initial behavior tree of the first unmanned equipment when the task completion degree is greater than a preset completion threshold; obtaining similar unmanned equipment when the task completion degree is less than or equal to the preset completion threshold; selecting collaborative unmanned equipment, updating the parallel node weights of the first unmanned equipment and the collaborative unmanned equipment, and performing collaborative scheduling of the unmanned equipment cluster. This solves the technical problems of insufficient flexibility, low collaborative efficiency, and poor environmental adaptability in existing unmanned equipment scheduling methods, leading to resource conflicts or unreasonable task allocation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of equipment scheduling, in particular to a behavior tree-based unmanned equipment cluster cooperative scheduling method and system. BACKGROUND

[0002] With the rapid development of artificial intelligence technology and unmanned equipment technology, unmanned equipment clusters are increasingly widely used in complex task scenarios. Unmanned equipment cluster cooperative scheduling, as a core link to achieve efficient cluster operation, directly affects the efficiency and success rate of task execution.

[0003] However, traditional unmanned equipment scheduling methods mostly rely on preset rules or centralized control algorithms. In the face of dynamically changing task environments, complex and diverse task targets, and large-scale equipment clusters, there are often problems such as insufficient scheduling flexibility, low cooperative efficiency, poor environmental adaptability, and scheduling decision delays. In addition, existing scheduling methods do not intuitively describe the cooperative behavior between unmanned equipment, making it difficult to accurately model the complex cooperative relationship between equipment, resulting in resource conflicts or unreasonable task allocation during cooperative task execution. SUMMARY

[0004] The embodiments of the present application provide a behavior tree-based unmanned equipment cluster cooperative scheduling method and system, which solves the technical problems of insufficient flexibility, low cooperative efficiency, poor environmental adaptability, and resulting resource conflicts or unreasonable task allocation of existing unmanned equipment scheduling methods.

[0005] The technical solution of the present application to solve the above technical problems is as follows: In a first aspect, the present application provides a behavior tree-based unmanned equipment cluster cooperative scheduling method, which comprises: Collecting equipment categories and basic state parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic state parameters include equipment state and task state; Randomly selecting one of the unmanned equipment to obtain a first unmanned equipment, and based on the task target of the first unmanned equipment, obtaining an initial behavior tree of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node; Based on the equipment category and the task state of the first unmanned equipment, the root node is updated to obtain an updated selection weight; Based on the equipment state and environmental parameters of the first unmanned equipment, task completion degree is evaluated. When the task completion degree is greater than a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated to obtain an updated behavior tree, and the scheduling of the first unmanned equipment is performed; When the task completion rate is less than or equal to the preset completion threshold, an unmanned equipment similar to the task target and sequence node of the first unmanned equipment is obtained as a similar unmanned equipment. Based on the task completion rate of the similar unmanned equipment, collaborative unmanned equipment is selected and analyzed for cooperation. The parallel node weights of the first unmanned equipment and the collaborative unmanned equipment are updated, the updated behavior tree is obtained, and the unmanned equipment cluster is coordinated and scheduled.

[0006] Secondly, this application provides a behavior tree-based collaborative scheduling system for unmanned equipment clusters, including: The parameter acquisition module is used to collect the equipment category and basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status. The behavior tree acquisition module is used to randomly select one of the unmanned equipments, acquire the first unmanned equipment, and acquire the initial behavior tree of the first unmanned equipment based on the mission objective of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node. The weight update module is used to update the weight of the root node based on the equipment category and the mission status of the first unmanned equipment, and obtain the update selection weight. The behavior tree update module is used to evaluate the task completion rate based on the equipment status and environmental parameters of the first unmanned equipment. When the task completion rate is greater than a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated to obtain the updated behavior tree and the first unmanned equipment is scheduled. The similar unmanned equipment acquisition module is used to acquire unmanned equipment similar to the task target and sequence node of the first unmanned equipment when the task completion degree is less than or equal to the preset completion threshold, and to identify them as similar unmanned equipment. The equipment scheduling module is used to filter and obtain collaborative unmanned equipment based on the task completion degree of the similar unmanned equipment, perform assistance analysis, update the parallel node weights of the first unmanned equipment and the collaborative unmanned equipment, obtain the updated behavior tree, and perform collaborative scheduling of the unmanned equipment cluster.

[0007] This application provides one or more technical solutions, which have at least the following technical effects or advantages: This application provides a behavior tree-based collaborative scheduling method and system for unmanned equipment clusters. First, it collects the equipment category and basic state parameters (including equipment status and task status) of each unmanned equipment in the cluster, providing fundamental data support for subsequent scheduling decisions. Second, it randomly selects one unmanned equipment as the first unmanned equipment and constructs an initial behavior tree containing a root node, sequence nodes, and parallel nodes based on its task objectives. This behavior tree serves as the basic framework for scheduling, reflecting the task execution logic of the equipment. Then, combining the equipment category and task status of the first unmanned equipment, it updates the weights of the root node to obtain updated selection weights, ensuring that the selection of the root node better aligns with the priority of equipment characteristics and task requirements. Next, it evaluates the task completion rate based on the equipment status and environmental parameters of the first unmanned equipment, ensuring the selection of the equipment with the best collaborative effect. Finally, it corrects the initial weights of the parallel nodes based on the collaborative effect parameters, and uses the updated parallel node weights and updated selection weights to update the initial behavior trees of the first unmanned equipment and the collaborating unmanned equipment, forming an updated behavior tree, thus achieving parallel collaborative scheduling of the unmanned equipment cluster.

[0008] Through the above technical solutions, this application effectively solves the problems of flexibility, collaborative efficiency and environmental adaptability of traditional scheduling methods in dynamic environments and complex tasks, avoids resource conflicts and unreasonable task allocation, and improves the overall task execution efficiency and success rate of the cluster. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart illustrating the behavior tree-based collaborative scheduling method for unmanned equipment clusters provided in this application embodiment; Figure 2 This is a schematic diagram of the structure of the behavior tree-based unmanned equipment cluster collaborative scheduling system provided in the embodiments of this application.

[0011] The components represented by each number in the attached diagram are explained below: Parameter acquisition module 11, behavior tree acquisition module 12, weight update module 13, behavior tree update module 14, similar unmanned equipment acquisition module 15, and equipment scheduling module 16. Detailed Implementation

[0012] This application provides a behavior tree-based collaborative scheduling method and system for unmanned equipment clusters, which addresses the technical problems of insufficient flexibility, low collaborative efficiency, and poor environmental adaptability in existing unmanned equipment scheduling methods, leading to resource conflicts or unreasonable task allocation.

[0013] Example 1, as Figure 1 As shown in the embodiments of this application, a collaborative scheduling method for unmanned equipment clusters based on behavior trees is provided, including: S10: Collect the equipment category and basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status. In this embodiment, data is collected from multiple unmanned devices in the unmanned equipment cluster, including equipment category information for each device, such as drones, unmanned vehicles, unmanned boats, or other types of unmanned equipment. Different equipment categories correspond to different functional characteristics and task execution capabilities. Simultaneously, basic status parameters are collected, including equipment status and task status.

[0014] Specifically, step S10 in the method includes: The equipment status of the unmanned equipment is collected, including remaining battery life, current location, and available payload. The mission status is obtained through the mission configuration of the unmanned equipment, and the mission status includes the mission area and mission priority.

[0015] In this embodiment of the application, firstly, the equipment status of the unmanned equipment is collected, specifically including remaining range, that is, the remaining range or distance that the unmanned equipment can continue to perform the task; current location, which is the real-time coordinate information obtained through positioning systems such as GPS and Beidou; and available load, which refers to the remaining capacity or weight of the task equipment, materials, etc. that the unmanned equipment can currently carry.

[0016] Secondly, the mission status is obtained through the mission configuration information of the unmanned equipment. The mission status specifically includes the mission area, which is the geographical range or spatial area where the unmanned equipment is assigned to perform the mission; and the mission priority, which is used to characterize the importance of the mission in the overall mission sequence. It is usually expressed in numerical or hierarchical form, and high-priority missions should be given priority in execution.

[0017] Specifically, by collecting data on equipment status and mission status, we can understand the current capabilities and mission requirements of unmanned equipment.

[0018] S20: Randomly select one of the unmanned equipment to obtain the first unmanned equipment, and based on the mission objective of the first unmanned equipment, obtain the initial behavior tree of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node and a parallel node; In this embodiment, an unmanned equipment is randomly selected from the unmanned equipment cluster and designated as the first unmanned equipment. Based on the specific task objectives undertaken by the first unmanned equipment, an initial behavior tree is constructed. This initial behavior tree is a structured description of the task execution logic of the first unmanned equipment. The root node serves as the starting point of the behavior tree, representing the initiation of the entire task objective and the highest-level decision. Sequence nodes represent a series of subtasks that need to be executed sequentially in a specific order. The execution of the next subtask will only begin after the previous subtask is successfully completed. Parallel nodes allow multiple subtasks under them to be processed simultaneously to improve the efficiency of task execution.

[0019] Specifically, step S20 in the method includes: Randomly select any unmanned equipment from the unmanned equipment cluster as the first unmanned equipment; Obtain the mission objective of the first unmanned equipment, and generate an initial behavior tree based on the mission objective and mission priority, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node.

[0020] In this embodiment of the application, firstly, any unmanned equipment is selected from the unmanned equipment cluster by random sampling and marked as the first unmanned equipment. This serves as the starting point for collaborative scheduling analysis, avoiding scheduling bias that may be caused by fixed selection rules.

[0021] Secondly, the mission objectives of the first unmanned equipment are analyzed to clarify its core tasks and specific indicators. Simultaneously, based on the task priority corresponding to the mission objective and the logical relationship of task execution, an initial behavior tree is generated. The root node of this initial behavior tree corresponds to the highest-level mission objective decision. Sequence nodes decompose the task into several sub-task units that need to be executed sequentially based on the task dependencies. For example, in a regional patrol mission, sequence nodes may include sub-tasks executed in sequence such as "takeoff," "heading to the target area," "patrolling along a predetermined route," and "returning." Parallel nodes are used to integrate sub-tasks that can be performed synchronously. For example, during patrol, "real-time image acquisition" and "environmental data monitoring" can be executed simultaneously as sub-tasks under parallel nodes, thereby improving mission processing efficiency.

[0022] By introducing task priorities, the execution order or resource allocation weight of subtasks within sequence nodes and parallel nodes can be initially set during the initial behavior tree construction, ensuring that high-priority subtasks are processed first.

[0023] S30: Based on the equipment category and mission status of the first unmanned equipment, update the weight of the root node and obtain the update selection weight; In this embodiment of the application, since different equipment categories have different inherent capabilities and task adaptability, the root node of the initial behavior tree is updated by combining the equipment category and task status of the first unmanned equipment to obtain the updated selection weight. By combining the capability weight brought by the equipment category and the priority weight in the task status, the dynamic update of the selection weight of the root node is realized, so that the root node can more accurately reflect the importance of the current task objective and the execution capability of the equipment.

[0024] Specifically, step S30 in the method includes: Map the equipment categories and determine the initial weight of the root node; Calculate the product of the task priority coefficient in the task state and the initial weight of the root node, and perform normalization processing to obtain the updated selection weight.

[0025] In this embodiment of the application, firstly, a mapping relationship between equipment category and initial weight of root node is established. For example, for UAVs with complex environment perception and high-precision execution capabilities, the initial weight of its root node can be set to a higher value, such as 0.8, while for ground unmanned vehicles with relatively simple functions, the initial weight can be set to a medium value, such as 0.5. The specific mapping rules can be pre-configured according to the actual application scenario and equipment characteristics of the unmanned equipment cluster.

[0026] Secondly, extract the task priority from the task status and convert it into a task priority coefficient. For example, the task priority is divided into three levels: high, medium, and low, corresponding to coefficients of 1.2, 1.0, and 0.8, respectively. Then, calculate the product of the initial weight of the root node and the task priority coefficient to obtain the temporary weight value of the root node.

[0027] Finally, the temporary weight value is normalized by dividing it by the maximum value among all possible root node weight values, so that the range of the update selection weight is limited to [0,1]. Through this process, the update selection weight of the root node can comprehensively reflect the inherent capabilities of the equipment and the urgency and importance of the current mission.

[0028] For example, if the initial weight of the root node corresponding to the equipment category of a certain UAV is 0.8, and its task priority is "high" with a coefficient of 1.2, then the temporary weight value is 0.8 × 1.2 = 0.96. Assuming that the maximum value among all possible root node weight values ​​is 1.2, then the updated selection weight is 0.96 / 1.2 = 0.8. This value serves as a reference for the root node in subsequent behavior tree decisions, guiding the initiation priority of the task objective.

[0029] S40: Based on the equipment status and environmental parameters of the first unmanned equipment, perform task completion evaluation. When the task completion is greater than a preset completion threshold, update the initial behavior tree of the first unmanned equipment, obtain the updated behavior tree, and schedule the first unmanned equipment. In this embodiment, before scheduling the first unmanned equipment, its own equipment status and environmental parameters are considered to assess its ability to independently complete the task, i.e., the task completion rate. The environmental parameters include information such as the terrain features, weather conditions, and the distribution of potential obstacles in the task execution area.

[0030] The task completion assessment, based on the equipment status and environmental parameters of the first unmanned equipment, includes: Based on the remaining endurance, current location, and nearest boundary of the mission area in the equipment status, obtain the location execution capability parameters; Based on the available load in the equipment status parameters and the mission objective, obtain the target completion capability parameters; Based on the environmental parameters, the mission objective, and the equipment category, historical environmental risks are obtained; The task completion rate is obtained by weighting the location execution capability parameters, the target completion capability parameters, and the historical environmental risks.

[0031] In this embodiment, the position execution capability parameter is first calculated, specifically by evaluating the distance between the current position of the first unmanned equipment and the nearest boundary of the mission area, combined with the maximum movement distance that can be supported by the remaining range.

[0032] For example, if the current location is 5 kilometers from the nearest boundary of the mission area, and the remaining battery life supports a movement of 10 kilometers, then the location execution capability parameter can be expressed as follows: After normalization, it can be set to 1. When the ratio is greater than 1, it is taken as 1, indicating that the unmanned equipment can easily reach the mission area.

[0033] Secondly, the target completion capability parameter is determined based on the ratio of available load to the load required by the mission target. For example, if the mission target requires a 2kg load and the available load is 3kg, then the target completion capability parameter is 3 / 2=1.5, which can be set to 1 after normalization. When the ratio is greater than 1, it is taken as 1, reflecting the sufficiency of the equipment to carry the mission equipment or materials.

[0034] Then, historical environmental risk is obtained by analyzing similar environmental parameters in historical execution data, such as terrain complexity, wind force level, and obstacle density, as well as the mission failure rate or abnormal event occurrence rate of this equipment category. For example, in a mountainous terrain environment with a wind force of level 5, the historical mission failure rate of the drone is 15%, and the historical environmental risk value is 0.15.

[0035] Finally, the location execution capability parameters, target completion capability parameters, and historical environmental risks are weighted and calculated to obtain the task completion rate. The weighting coefficients are pre-set based on the key influencing factors of the task objectives.

[0036] For example, a weighted calculation is performed with the location execution capability parameter weight of 0.4, the target completion capability parameter weight of 0.4, and the historical environmental risk weight of 0.2, resulting in a weighted calculation of location execution capability parameter 0.5, target completion capability parameter 1, and historical environmental risk 0.15. If the preset completion threshold is 0.7, then the task completion rate of 0.97 is greater than the threshold. At this time, the initial behavior tree is updated based on the equipment status and environmental parameters.

[0037] Specifically, when the task completion rate is greater than a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated, the updated behavior tree is obtained, and the first unmanned equipment is scheduled, including: Based on the task objective, a preset completion threshold is obtained through mapping. When the task completion rate is greater than the preset completion threshold, the parallel node adaptation coefficient is calculated based on the equipment status and the environmental parameters, the selection weight of the parallel node is updated, and the updated parallel weight is obtained. The initial behavior tree is updated by combining the update selection weight and the update parallel weight to obtain the updated behavior tree, and the first unmanned equipment is scheduled.

[0038] In this embodiment, firstly, based on the type and complexity of the mission objective, a corresponding preset completion threshold is obtained from a preset threshold database. For example, for a simple fixed-point reconnaissance mission, the preset completion threshold can be set to 0.6, while for a complex multi-target collaborative search and rescue mission, the preset completion threshold can be increased to 0.8 to ensure that the equipment has a high independent completion capability.

[0039] Secondly, when the task completion rate is greater than the preset completion threshold, it indicates that the first unmanned equipment has the basic conditions to independently execute the task in the current state. At this time, the parallel node adaptation coefficient is calculated in combination with its equipment status and environmental parameters to adjust the selection weight of the parallel nodes.

[0040] Specifically, the calculation of the parallel node fitness coefficient comprehensively considers the remaining endurance's ability to support the number of parallel tasks, the impact of environmental parameters on the execution accuracy of parallel tasks, and the available load's capacity to support the equipment required for parallel tasks. For example, "parallel node fitness coefficient = position execution capability parameter × target completion capability parameter × (1 - historical environmental risk)".

[0041] Then, the fitness coefficient of the parallel node is multiplied by the original base weight of the parallel node to obtain the updated parallel weight, which dynamically adjusts the execution priority and resource allocation ratio of each subtask under the parallel node. Finally, the updated selection weight obtained in step S30 is used as the decision basis for the root node. Combined with the updated parallel weight, the node structure and weight configuration of the initial behavior tree are updated as a whole to generate an updated behavior tree that can adapt to the current equipment status and environmental conditions. Based on the updated behavior tree, the first unmanned equipment is scheduled to execute tasks according to the optimized logic.

[0042] S50: When the task completion rate is less than or equal to the preset completion threshold, obtain unmanned equipment similar to the task target and sequence node of the first unmanned equipment, and use it as a similar unmanned equipment. In this embodiment of the application, when the task completion rate does not reach the preset completion threshold, it indicates that the first unmanned equipment has insufficient ability to complete the task independently. At this time, other unmanned equipment that can provide collaborative support are searched from the cluster.

[0043] Specifically, step S50 in the method includes: When the task completion rate is less than or equal to the preset completion threshold, unmanned equipment is screened in the unmanned equipment cluster to obtain unmanned equipment with task target similarity greater than the target similarity threshold and sequence node similarity greater than the node similarity threshold, which are then regarded as similar unmanned equipment.

[0044] In this embodiment, the calculation method for task target similarity is first clarified. Core elements of the first unmanned equipment's task target, such as task type, task area range, and key task indicators, are extracted and matched with the task targets of other unmanned equipment in the cluster. For example, if the task target of the first unmanned equipment is "to conduct high-definition image reconnaissance of area A," then the core elements include "reconnaissance," "area A," and "high-definition image." If the task targets of other unmanned equipment contain the same or highly similar elements, then the task target similarity is high. The similarity value of the task target text is calculated using a cosine similarity algorithm. For example, a target similarity threshold of 0.7 is set. When the similarity value is greater than 0.7, the task targets are considered similar.

[0045] Specifically, cosine similarity is an algorithm that measures similarity by calculating the cosine of the angle between two vectors. In task target similarity calculation, the task target text is first converted into vector form, which can be achieved through the bag-of-words model or word embedding techniques, such as Word2Vec. Each task target text is represented as a high-dimensional vector, and each dimension of the vector corresponds to a feature word and its weight. Then, the cosine value of the two vectors is calculated using the formula: , where A and B are the vectors corresponding to the two task target texts, A·B represents the dot product of the vectors, and ||A|| and ||B|| represent the magnitudes of the vectors, respectively. The closer the cosine value is to 1, the higher the similarity between the two task target texts.

[0046] Secondly, sequence node similarity is determined by comparing the matching degree of the subtask sequences of sequence nodes in the initial behavior tree of the first unmanned equipment with the subtask sequences of sequence nodes in the behavior trees of other unmanned equipment. For example, the edit distance algorithm is used to calculate the difference between two subtask sequences; the smaller the difference, the higher the similarity. A node similarity threshold of 0.6 is set, and when the sequence node similarity is greater than 0.6, the sequence nodes are considered similar. By simultaneously satisfying the dual screening conditions of task target similarity and sequence node similarity, similar unmanned equipment that can form synergy with the first unmanned equipment in terms of task target and execution process is selected from the unmanned equipment cluster.

[0047] The edit distance algorithm measures the difference between sequences by calculating the minimum number of edit operations required to transform one subtask sequence into another. The smaller the edit distance, the higher the similarity between the sequence nodes. For example, if the sequence node subtask of the first unmanned equipment is "takeoff → proceed to target area → patrol → return", and the sequence node subtask of the other unmanned equipment is "takeoff → proceed to target area → search → return", and the edit distance between the two is 1 (replacing "patrol" with "search"), then if the difference threshold is set to 1, it can be determined that the sequence nodes of the two are highly similar.

[0048] S60: Based on the task completion rate of the similar unmanned equipment, filter and obtain collaborative unmanned equipment, perform assistance analysis, update the parallel node weights of the first unmanned equipment and the collaborative unmanned equipment, obtain the updated behavior tree, and perform collaborative scheduling of the unmanned equipment cluster.

[0049] In this embodiment, after selecting similar unmanned equipment, their task completion rate is further evaluated to determine whether they have the ability to perform tasks collaboratively. Specifically, based on the equipment status and environmental parameters of the similar unmanned equipment, the same task completion rate evaluation method as the first unmanned equipment is used to calculate the task completion rate of each similar unmanned equipment, and then collaborative unmanned equipment is selected.

[0050] After acquiring the collaborative unmanned equipment, a collaborative analysis is performed, which includes calculating the collaborative capability matching degree and evaluating the collaborative cost. The parallel node weights of the first unmanned equipment and the collaborative unmanned equipment are then updated to form an update behavior tree, and collaborative scheduling of the unmanned equipment cluster is carried out.

[0051] Specifically, step S60 in the method includes: The equipment with the highest mission completion rate among the similar unmanned equipment is selected as the collaborative unmanned equipment; A collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters; When the collaborative effect parameter is less than or equal to the collaborative effect threshold, the equipment with the highest task completion rate among the similar unmanned equipment (excluding the currently selected collaborative unmanned equipment) is selected as the updated collaborative unmanned equipment, and the collaborative analysis is iterated until an updated collaborative unmanned equipment with a collaborative effect parameter greater than the collaborative effect threshold is obtained as the collaborative unmanned equipment. When the collaborative effect parameter is greater than the collaborative effect threshold, the initial weight of the parallel node is adjusted by combining the collaborative effect parameter to obtain the updated weight of the parallel node. The initial behavior tree of the first unmanned equipment is updated using the updated parallel node weights and the updated selection weights to form an updated behavior tree; The updated behavior tree is used to perform parallel collaborative scheduling of the first unmanned equipment and the cooperative unmanned equipment.

[0052] In this embodiment of the application, firstly, since equipment with high task completion rate usually has better independent execution capability and can provide more reliable support for collaborative tasks, the equipment with the highest task completion rate among similar unmanned equipment is selected as the candidate for collaborative unmanned equipment. For example, if the task completion rate of the first unmanned equipment is 0.65 and the preset completion threshold is 0.7, while the task completion rate of unmanned vehicle A among similar unmanned equipment is 0.85 and the task completion rate of unmanned boat B is 0.78, then unmanned vehicle A is selected as the initial candidate for collaborative unmanned equipment.

[0053] Secondly, a synergy analysis is conducted between the first unmanned equipment and the candidate collaborative unmanned equipment to calculate the synergy capability matching degree and evaluate the synergy cost, thereby obtaining synergy effect parameters. The synergy capability matching degree mainly considers the complementarity between the two in terms of equipment functions and mission execution capabilities.

[0054] Then, the calculated collaborative effect parameters are compared with the preset collaborative effect threshold. If the collaborative effect parameters are less than or equal to the collaborative effect threshold, it means that the collaborative effect between the current candidate collaborative unmanned equipment and the first unmanned equipment has not met expectations and needs to be reselected.

[0055] For example, if the collaborative effect parameter of the aforementioned unmanned vehicle A is 0.595, which is less than the threshold of 0.6, then unmanned vehicle A is excluded, and unmanned vessel B with the highest task completion rate is selected as the updated collaborative unmanned equipment, and the collaborative analysis is performed again. Assuming the collaborative capability matching degree between unmanned vessel B and the first unmanned aerial vehicle is 0.75, and the collaborative cost assessment is 0.2, then the collaborative effect parameter is... This value equals the collaborative effect threshold, at which point unmanned vessel B can be identified as the final collaborative unmanned equipment. If the recalculated collaborative effect parameters still do not meet the requirements, the selection continues iteratively until a collaborative unmanned equipment with collaborative effect parameters greater than the collaborative effect threshold is found. If no collaborative unmanned equipment with collaborative effect parameters greater than the collaborative effect threshold can be found through continuous iteration, a prompt is issued and manual intervention is requested.

[0056] When the collaborative effect parameter is greater than the collaborative effect threshold, it indicates that the first unmanned equipment and the collaborative unmanned equipment have a good collaborative foundation. At this time, the initial weights of the parallel nodes are corrected in combination with the collaborative effect parameter to obtain updated parallel node weights.

[0057] For example, if the base weight of a parallel node in the initial behavior tree of the first unmanned equipment is 0.6 and the cooperative effect parameter is 0.7, then the updated parallel node weight = This indicates that in collaborative tasks, the execution priority and resource allocation of subtasks under this parallel node will be adjusted according to the collaborative effect in order to optimize the overall collaborative efficiency.

[0058] Furthermore, the initial behavior tree of the first unmanned equipment is updated using the updated parallel node weights and the updated selection weights obtained in step S30. The updated behavior tree not only reflects the equipment's own capabilities and task priorities, but also incorporates the influence of cooperating equipment, enabling the behavior tree to guide the first unmanned equipment to perform tasks more rationally in a cooperative environment.

[0059] Specifically, when the behavior tree is executed in a loop, the root node prioritizes each sequence node according to the update selection weight, and executes the sequence node with the higher weight first; while within the parallel node, the execution order and resource allocation ratio of each subtask are determined by updating the weight of the parallel node, and the subtask with the higher weight will get more computing resources and execution time slices.

[0060] Finally, based on the updated behavior tree, the first unmanned equipment and the collaborative unmanned equipment are scheduled in parallel. For example, the drone is responsible for large-scale aerial reconnaissance and real-time image data transmission, while the unmanned vessel is responsible for sampling or material delivery in specific waters based on the information provided by the drone. The two achieve seamless connection of task flow and efficient use of resources through the updated behavior tree, and jointly complete complex task objectives.

[0061] Furthermore, a collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters, including: Obtain the first task area of ​​the first unmanned equipment and the collaborative task area of ​​the collaborative unmanned equipment; Calculate the ratio of the intersection area of ​​the first task region and the collaborative task region to the area of ​​the first task region to obtain the first collaborative area parameter; Obtain the parallel nodes of the first unmanned equipment and the cooperative unmanned equipment, calculate the ratio of the weight of the parallel node of the first unmanned equipment to the weight of the same parallel node in the cooperative unmanned equipment, and obtain the similarity of the first parallel node. Obtain the ratio of the remaining range of the collaborative unmanned equipment after completing the root node and sequence node to the standard range, and obtain the range redundancy parameter; Obtain the ratio of the available load to the standard load for the collaborative unmanned equipment to complete the root node and sequence nodes, and obtain the load redundancy parameter; The first collaborative area parameter, the first parallel node similarity, the endurance redundancy parameter, and the load redundancy parameter are weighted and summed to obtain the collaborative effect parameter.

[0062] In this embodiment of the application, firstly, a first collaborative area parameter is calculated, which reflects the degree of overlap between the mission areas of the first unmanned equipment and the collaborative unmanned equipment.

[0063] For example, the mission area of ​​the first unmanned equipment is a circular area with a radius of 5 kilometers, covering an area of ​​approximately 78.5 square kilometers. The collaborative mission area of ​​the cooperating unmanned equipment is a circular area with a radius of 3 kilometers that partially overlaps with the first unmanned equipment. The calculated intersection area is 20 square kilometers. Therefore, the first collaborative area parameter = The larger the value of this parameter, the more the task areas of the two overlap, the better the geographical foundation for collaboration, and the greater the convenience of information sharing and task coordination.

[0064] Secondly, the similarity of the first parallel nodes is calculated by comparing the ratio of the weights of the parallel nodes in the two systems to measure the similarity of the task execution logic. Assuming the weight of the parallel node "image acquisition" in the first unmanned equipment is 0.3, and the weight of the same parallel node "image acquisition" in the collaborative unmanned equipment is 0.4, then the similarity of the first parallel nodes = The closer the ratio is to 1, the more similar the two are in terms of resource allocation and execution priority on the parallel node. When cooperating, it is easier to reach a consensus on the understanding of task objectives and execution pace, thus reducing coordination conflicts.

[0065] Secondly, the endurance redundancy parameter is calculated as the ratio of the remaining endurance required for the collaborative unmanned equipment to complete the tasks of the current root node and sequence nodes to the standard endurance. If the standard endurance of the collaborative unmanned equipment is 120 minutes, and it is estimated that it will take another 40 minutes to complete the current task sequence, then the remaining endurance is 80 minutes, and the endurance redundancy parameter = This parameter reflects the sufficiency of the collaborative unmanned equipment's own endurance after assisting the primary unmanned equipment in completing its mission. The higher the value, the stronger the sustainability of the collaborative mission and the lower the risk of collaborative interruption due to insufficient endurance.

[0066] Then, the load redundancy parameter is the ratio of the available load of the collaborative unmanned equipment to the standard load. For example, if the standard load of the collaborative unmanned equipment is 50 kg, and the currently loaded mission equipment weighs 30 kg, then the available load is 20 kg, and the load redundancy parameter = This parameter reflects the ability of collaborative unmanned equipment to carry additional equipment or supplies required for collaborative tasks. The larger the value, the greater the potential for collaborative support, such as carrying more sensors or supplies.

[0067] Finally, the first collaborative area parameter, the first parallel node similarity parameter, the endurance redundancy parameter, and the load redundancy parameter are weighted and summed to obtain the collaborative effect parameter. The weights are adjusted according to the different task types and collaborative requirements.

[0068] For example, in collaborative tasks emphasizing regional coverage, the weight of the first collaborative area parameter can be set to 0.3; in scenarios focusing on the consistency of task execution logic, the weight of the first parallel node similarity can be set to 0.3; the endurance redundancy and load redundancy parameters are assigned weights of 0.2 and 0.2 respectively. Using the above example data, the collaborative effect parameter = By comparing with the synergistic effect threshold, it can be determined whether the current synergistic combination meets the requirements.

[0069] In summary, compared to existing technologies, this application constructs a complete collaborative scheduling mechanism for unmanned equipment clusters within a behavior tree framework. First, through dual screening using task objective similarity and sequence node similarity, it identifies similar unmanned equipment with collaborative potential in both task objectives and execution processes, avoiding resource waste and inefficiency caused by blind collaboration. Second, it introduces task completion assessment and combines it with collaborative capability matching and collaborative cost assessment for assistance analysis, ensuring that the selected collaborative unmanned equipment possesses actual collaborative execution capabilities and can participate in tasks with superior collaborative effects, thus improving the reliability and effectiveness of collaboration. Third, by updating the weights of parallel nodes to dynamically adjust the behavior tree, it enables the behavior tree to not only reflect the task execution logic of individual equipment but also incorporate the influence of collaborative equipment, achieving dynamic optimization of the behavior tree. This guides unmanned equipment to allocate resources and adjust task priorities more rationally in a collaborative environment, ultimately achieving efficient parallel collaborative scheduling of unmanned equipment clusters.

[0070] In summary, the embodiments of this application have at least the following technical effects: This application provides a behavior tree-based collaborative scheduling method for unmanned equipment clusters. First, it collects the equipment category and basic state parameters (including equipment status and task status) of each unmanned equipment in the cluster. Second, it randomly selects one unmanned equipment as the first unmanned equipment and constructs an initial behavior tree containing a root node, sequence nodes, and parallel nodes based on its task objective. This behavior tree serves as the basic framework for scheduling. Then, it updates the weights of the root node based on the equipment category and task status of the first unmanned equipment to obtain update selection weights. Next, it evaluates the task completion rate based on the equipment status and environmental parameters of the first unmanned equipment. Finally, it corrects the initial weights of the parallel nodes based on the collaborative effect parameters and uses the updated parallel node weights and update selection weights to update the initial behavior trees of the first unmanned equipment and the collaborating unmanned equipment, forming an updated behavior tree.

[0071] Through the above technical solutions, this application effectively solves the problems of flexibility, collaborative efficiency and environmental adaptability of traditional scheduling methods in dynamic environments and complex tasks, avoids resource conflicts and unreasonable task allocation, and improves the overall task execution efficiency and success rate of the cluster.

[0072] Example 2, as Figure 2 As shown, based on the same inventive concept as the behavior tree-based unmanned equipment cluster collaborative scheduling method provided in Embodiment 1, this application also provides a behavior tree-based unmanned equipment cluster collaborative scheduling system, including: The parameter acquisition module 11 is used to acquire the equipment category and basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status. The behavior tree acquisition module 12 is used to randomly select one of the unmanned equipments, acquire the first unmanned equipment, and acquire the initial behavior tree of the first unmanned equipment based on the mission objective of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node. The weight update module 13 is used to update the weight of the root node based on the equipment category and the mission status of the first unmanned equipment, and obtain the update selection weight. The behavior tree update module 14 is used to evaluate the task completion rate based on the equipment status and environmental parameters of the first unmanned equipment. When the task completion rate is greater than a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated to obtain the updated behavior tree and the first unmanned equipment is scheduled. The similar unmanned equipment acquisition module 15 is used to acquire unmanned equipment similar to the task target and the sequence node of the first unmanned equipment when the task completion degree is less than or equal to the preset completion threshold, and to use it as similar unmanned equipment. The equipment scheduling module 16 is used to filter and obtain collaborative unmanned equipment based on the task completion degree of the similar unmanned equipment, perform assistance analysis, update the parallel node weights of the first unmanned equipment and the collaborative unmanned equipment, obtain the updated behavior tree, and perform collaborative scheduling of the unmanned equipment cluster.

[0073] In one embodiment, the parameter acquisition module 11 is specifically used for: The equipment status of the unmanned equipment is collected, including remaining battery life, current location, and available payload. The mission status is obtained through the mission configuration of the unmanned equipment, and the mission status includes the mission area and mission priority.

[0074] In one embodiment, the behavior tree acquisition module 12 is specifically used for: Randomly select any unmanned equipment from the unmanned equipment cluster as the first unmanned equipment; Obtain the mission objective of the first unmanned equipment, and generate an initial behavior tree based on the mission objective and mission priority, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node.

[0075] In one embodiment, the weight update module 13 is specifically used for: Map the equipment categories and determine the initial weight of the root node; Calculate the product of the task priority coefficient in the task state and the initial weight of the root node, and perform normalization processing to obtain the updated selection weight.

[0076] Furthermore, in one embodiment of the application, based on the equipment status and environmental parameters of the first unmanned equipment, a mission completion assessment is performed, including: Based on the remaining endurance, current location, and nearest boundary of the mission area in the equipment status, obtain the location execution capability parameters; Based on the available load in the equipment status parameters and the mission objective, obtain the target completion capability parameters; Based on the environmental parameters, the mission objective, and the equipment category, historical environmental risks are obtained; The task completion rate is obtained by weighting the location execution capability parameters, the target completion capability parameters, and the historical environmental risks.

[0077] Furthermore, when the task completion rate exceeds a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated to obtain the updated behavior tree, and the first unmanned equipment is scheduled, including: Based on the task objective, a preset completion threshold is obtained through mapping. When the task completion rate is greater than the preset completion threshold, the parallel node adaptation coefficient is calculated based on the equipment status and the environmental parameters, the selection weight of the parallel node is updated, and the updated parallel weight is obtained. The initial behavior tree is updated by combining the update selection weight and the update parallel weight to obtain the updated behavior tree, and the first unmanned equipment is scheduled.

[0078] In one embodiment, the similar unmanned equipment acquisition module 15 is specifically used for: When the task completion rate is less than or equal to the preset completion threshold, unmanned equipment is screened in the unmanned equipment cluster to obtain unmanned equipment with task target similarity greater than the target similarity threshold and sequence node similarity greater than the node similarity threshold, which are then regarded as similar unmanned equipment.

[0079] In one embodiment, the equipment scheduling module 16 is specifically used for: The equipment with the highest mission completion rate among the similar unmanned equipment is selected as the collaborative unmanned equipment; A collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters; When the collaborative effect parameter is less than or equal to the collaborative effect threshold, the equipment with the highest task completion rate among the similar unmanned equipment (excluding the currently selected collaborative unmanned equipment) is selected as the updated collaborative unmanned equipment, and the collaborative analysis is iterated until an updated collaborative unmanned equipment with a collaborative effect parameter greater than the collaborative effect threshold is obtained as the collaborative unmanned equipment. When the collaborative effect parameter is greater than the collaborative effect threshold, the initial weight of the parallel node is adjusted by combining the collaborative effect parameter to obtain the updated weight of the parallel node. The initial behavior tree of the first unmanned equipment is updated using the updated parallel node weights and the updated selection weights to form an updated behavior tree; The updated behavior tree is used to perform parallel collaborative scheduling of the first unmanned equipment and the cooperative unmanned equipment.

[0080] Furthermore, a collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters, including: Obtain the first task area of ​​the first unmanned equipment and the collaborative task area of ​​the collaborative unmanned equipment; Calculate the ratio of the intersection area of ​​the first task region and the collaborative task region to the area of ​​the first task region to obtain the first collaborative area parameter; Obtain the parallel nodes of the first unmanned equipment and the cooperative unmanned equipment, calculate the ratio of the weight of the parallel node of the first unmanned equipment to the weight of the same parallel node in the cooperative unmanned equipment, and obtain the similarity of the first parallel node. Obtain the ratio of the remaining range of the collaborative unmanned equipment after completing the root node and sequence node to the standard range, and obtain the range redundancy parameter; Obtain the ratio of the available load to the standard load for the collaborative unmanned equipment to complete the root node and sequence nodes, and obtain the load redundancy parameter; The first collaborative area parameter, the first parallel node similarity, the endurance redundancy parameter, and the load redundancy parameter are weighted and summed to obtain the collaborative effect parameter.

Claims

1. A behavior tree-based collaborative scheduling method for unmanned equipment clusters, characterized in that, include: Collect equipment categories and basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status. Randomly select one of the unmanned equipment to obtain the first unmanned equipment, and based on the mission objective of the first unmanned equipment, obtain the initial behavior tree of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node and a parallel node; Based on the equipment category and mission status of the first unmanned equipment, the root node is weighted and the update selection weight is obtained. Based on the equipment status and environmental parameters of the first unmanned equipment, the task completion rate is evaluated. When the task completion rate is greater than the preset completion threshold, the initial behavior tree of the first unmanned equipment is updated, the updated behavior tree is obtained, and the first unmanned equipment is scheduled. When the task completion rate is less than or equal to the preset completion threshold, an unmanned equipment similar to the task target and sequence node of the first unmanned equipment is obtained as a similar unmanned equipment. Based on the task completion rate of the similar unmanned equipment, collaborative unmanned equipment is selected and analyzed for cooperation. The parallel node weights of the first unmanned equipment and the collaborative unmanned equipment are updated, the updated behavior tree is obtained, and the unmanned equipment cluster is coordinated and scheduled.

2. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, Collect basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status, including: The equipment status of the unmanned equipment is collected, including remaining battery life, current location, and available payload. The mission status is obtained through the mission configuration of the unmanned equipment, and the mission status includes the mission area and mission priority.

3. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, Randomly select one of the unmanned equipment to obtain a first unmanned equipment, and based on the mission objective of the first unmanned equipment, obtain an initial behavior tree for the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node, including: Randomly select any unmanned equipment from the unmanned equipment cluster as the first unmanned equipment; Obtain the mission objective of the first unmanned equipment, and generate an initial behavior tree based on the mission objective and mission priority, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node.

4. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, Based on the equipment category and mission status of the first unmanned equipment, the root node is weighted and updated to obtain the update selection weight, including: Map the equipment categories and determine the initial weight of the root node; Calculate the product of the task priority coefficient in the task state and the initial weight of the root node, and perform normalization processing to obtain the updated selection weight.

5. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, Based on the equipment status and environmental parameters of the first unmanned equipment, a mission completion assessment is performed, including: Based on the remaining endurance, current location, and nearest boundary of the mission area in the equipment status, obtain the location execution capability parameters; Based on the available load in the equipment status parameters and the mission objective, obtain the target completion capability parameters; Based on the environmental parameters, the mission objective, and the equipment category, historical environmental risks are obtained. The task completion rate is obtained by weighting the location execution capability parameters, the target completion capability parameters, and the historical environmental risks.

6. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, When the task completion rate exceeds a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated, the updated behavior tree is obtained, and the first unmanned equipment is scheduled, including: Based on the task objective, a preset completion threshold is obtained through mapping. When the task completion rate is greater than the preset completion threshold, the parallel node adaptation coefficient is calculated based on the equipment status and the environmental parameters, the selection weight of the parallel node is updated, and the updated parallel weight is obtained. The initial behavior tree is updated by combining the update selection weight and the update parallel weight to obtain the updated behavior tree, and the first unmanned equipment is scheduled.

7. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, When the task completion rate is less than or equal to the preset completion threshold, unmanned equipment similar to the task target and sequence node of the first unmanned equipment is acquired as similar unmanned equipment, including: When the task completion rate is less than or equal to the preset completion threshold, unmanned equipment is screened in the unmanned equipment cluster to obtain unmanned equipment with task target similarity greater than the target similarity threshold and sequence node similarity greater than the node similarity threshold, which are then regarded as similar unmanned equipment.

8. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 1, characterized in that, Based on the task completion rates of the similar unmanned equipment, collaborative unmanned equipment is selected and subjected to assistance analysis. The parallel node weights of the first unmanned equipment and the collaborative unmanned equipment are updated, the updated behavior tree is obtained, and collaborative scheduling of the unmanned equipment cluster is performed, including: The equipment with the highest mission completion rate among the similar unmanned equipment is selected as the collaborative unmanned equipment; A collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters; When the collaborative effect parameter is less than or equal to the collaborative effect threshold, the equipment with the highest task completion rate among the similar unmanned equipment (excluding the currently selected collaborative unmanned equipment) is selected as the updated collaborative unmanned equipment, and the collaborative analysis is iterated until an updated collaborative unmanned equipment with a collaborative effect parameter greater than the collaborative effect threshold is obtained as the collaborative unmanned equipment. When the collaborative effect parameter is greater than the collaborative effect threshold, the initial weight of the parallel node is adjusted by combining the collaborative effect parameter to obtain the updated weight of the parallel node. The initial behavior tree of the first unmanned equipment is updated using the updated parallel node weights and the updated selection weights to form an updated behavior tree; The updated behavior tree is used to perform parallel collaborative scheduling of the first unmanned equipment and the cooperative unmanned equipment.

9. The unmanned equipment cluster collaborative scheduling method based on behavior tree according to claim 8, characterized in that, A collaborative analysis is performed on the first unmanned equipment and the collaborative unmanned equipment to obtain collaborative effect parameters, including: Obtain the first task area of ​​the first unmanned equipment and the collaborative task area of ​​the collaborative unmanned equipment; Calculate the ratio of the intersection area of ​​the first task region and the collaborative task region to the area of ​​the first task region to obtain the first collaborative area parameter; Obtain the parallel nodes of the first unmanned equipment and the cooperative unmanned equipment, calculate the ratio of the weight of the parallel node of the first unmanned equipment to the weight of the same parallel node in the cooperative unmanned equipment, and obtain the similarity of the first parallel node. Obtain the ratio of the remaining range of the collaborative unmanned equipment after completing the root node and sequence node to the standard range, and obtain the range redundancy parameter; Obtain the ratio of the available load to the standard load for the collaborative unmanned equipment to complete the root node and sequence nodes, and obtain the load redundancy parameter; The first collaborative area parameter, the first parallel node similarity, the endurance redundancy parameter, and the load redundancy parameter are weighted and summed to obtain the collaborative effect parameter.

10. A behavior tree-based unmanned equipment cluster collaborative scheduling system, characterized in that, The method for executing the behavior tree-based collaborative scheduling method for unmanned equipment clusters as described in any one of claims 1-9 includes: The parameter acquisition module is used to collect the equipment category and basic status parameters of multiple unmanned equipment in the unmanned equipment cluster, wherein the basic status parameters include equipment status and mission status. The behavior tree acquisition module is used to randomly select one of the unmanned equipments, acquire the first unmanned equipment, and acquire the initial behavior tree of the first unmanned equipment based on the mission objective of the first unmanned equipment, wherein the initial behavior tree includes a root node, a sequence node, and a parallel node. The weight update module is used to update the weight of the root node based on the equipment category and the mission status of the first unmanned equipment, and obtain the update selection weight. The behavior tree update module is used to evaluate the task completion rate based on the equipment status and environmental parameters of the first unmanned equipment. When the task completion rate is greater than a preset completion threshold, the initial behavior tree of the first unmanned equipment is updated to obtain the updated behavior tree and the first unmanned equipment is scheduled. The similar unmanned equipment acquisition module is used to acquire unmanned equipment similar to the task target and sequence node of the first unmanned equipment when the task completion degree is less than or equal to the preset completion threshold, and to identify them as similar unmanned equipment. The equipment scheduling module is used to filter and obtain collaborative unmanned equipment based on the task completion degree of the similar unmanned equipment, perform assistance analysis, update the parallel node weights of the first unmanned equipment and the collaborative unmanned equipment, obtain the updated behavior tree, and perform collaborative scheduling of the unmanned equipment cluster.

Images