A robot behavior scheduling method, device and medium combining modular skills

Abstract

The present invention discloses a robot behavior scheduling method, device and medium combined with modular skills, and relates to the technical field related to robot scheduling management. The method includes: connecting to the SkiRobot platform and receiving task instructions; determining the target robot according to the application scenario task; disassembling the application scenario task to obtain N application subtasks; activating the robot behavior scheduling channel embedded in the SkiRobot platform; scheduling each skill module of the target robot and establishing a robot behavior scheduling tree; based on the SkiRobot platform, controlling the target robot to perform the application scenario task according to the robot behavior scheduling tree. The method solves the technical problem that it is difficult to conveniently and flexibly realize complex behavior combination and dynamic scheduling in the existing robot behavior scheduling, which leads to insufficient convenience and flexibility of robot behavior control and low scene adaptability, and achieves the technical effect of improving the convenience and flexibility of robot behavior control.

Description

Technical Field

[0001] The present application relates to the technical field related to robot scheduling management, and specifically to a robot behavior scheduling method, device and medium combined with modular skills. Background Art

[0002] In recent years, robotics has made significant progress in the fields of industrial automation, service robots, and special robots. The driving force behind this trend is the industry's continuous demand for improved efficiency, flexibility, and autonomy. In order to meet these needs, more and more robot operating systems have been developed and used, and ROS (Robot Operating System) and its subsequent version ROS2 have occupied an important position in them. As an important operating system in the field of robot development, ROS2 provides underlying communication, node management, and a variety of tools to help developers build complex robot systems. However, these functions are more focused on underlying communication and node management. For the construction and scheduling of complex behaviors, developers need to perform a lot of custom development and system integration. That is, although ROS2 can support communication between different nodes, it is not convenient enough to realize complex behavior combinations and dynamic scheduling, which affects the efficiency of task execution. In addition, the behavior scheduling and control optimization of task execution in some specific scenarios of ROS2 are not sophisticated enough, which increases the complexity of integration and the difficulty of implementation. In addition, the learning curve of ROS2 is relatively steep, and the learning cost and difficulty of getting started are high.

[0003] Therefore, in the current robot behavior scheduling related technologies, it is difficult to conveniently and flexibly realize complex behavior combinations and dynamic scheduling, which leads to technical problems such as insufficient convenience and flexibility of robot behavior control and low scene adaptability. Summary of the invention

[0004] The present application provides a robot behavior scheduling method, device and medium that combine modular skills, thereby solving the technical problems of the existing robot behavior scheduling, that is, the difficulty in conveniently and flexibly realizing complex behavior combination and dynamic scheduling, which in turn leads to the lack of convenience and flexibility of robot behavior control and low scene adaptability, thereby achieving the technical effect of improving the convenience and flexibility of robot behavior control.

[0005] The present application provides a robot behavior scheduling method combined with modular skills, the method comprising: connecting to a SkiRobot platform, receiving a task instruction, wherein the task instruction comprises an application scenario task, the SkiRobot platform is a robot control platform based on ROS2, the SkiRobot platform is communicatively connected to a plurality of control terminals of a plurality of robots, each robot comprising a plurality of skill modules; based on application scheduling bilateral constraints, performing application scheduling on the plurality of robots according to the application scenario task, and determining a target robot; disassembling the application scenario task, and obtaining N application subtasks, wherein N is a positive integer greater than 1; activating a robot behavior scheduling channel embedded in the SkiRobot platform, wherein the robot behavior scheduling channel comprises a skill behavior scheduling decision unit, a behavior scheduling response expectation constraint unit, and a behavior scheduling abnormality prediction operation and maintenance unit; based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit, and the behavior scheduling abnormality prediction operation and maintenance unit, scheduling each skill module of the target robot according to the N application subtasks, and establishing a robot behavior scheduling tree; based on the SkiRobot platform, controlling the target robot to perform the application scenario task according to the robot behavior scheduling tree.

[0006] In a possible implementation, based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, each skill module of the target robot is scheduled according to the N application subtasks, and a robot behavior scheduling tree is established, and the following processing is also performed: based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, each skill module is scheduled according to the N application subtasks respectively to obtain N task behavior scheduling strategies; the application scenario task is set as the scheduling tree root node; the N application subtasks are N scheduling branch nodes, and the N task behavior scheduling strategies are N scheduling leaf nodes; the logical relationship between the application scenario task, the N application subtasks and the N task behavior scheduling strategies is the scheduling tree topological relationship; based on the scheduling tree topological relationship, the scheduling tree root node, the N scheduling branch nodes and the N scheduling leaf nodes are topologically connected to generate the robot behavior scheduling tree.

[0007] In a possible implementation, based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, the various skill modules are scheduled respectively according to the N application subtasks to obtain N task behavior scheduling strategies, and the following processing is also performed: according to the N application subtasks, the nth application subtask is extracted, where n is a positive integer, 1≤n≤N; based on the nth application subtask, the various skill modules are associated with an analysis to determine the nth task-associated skill module; according to the SkiRobot platform, the basic information of the nth task-associated skill module is collected to obtain the nth skill module data; the nth application subtask and the nth skill module are associated. The data of the energy module is input into the skill behavior scheduling decision unit to obtain the nth skill behavior scheduling decision; the scheduling response expectation constraint is performed on the nth skill behavior scheduling decision according to the behavior scheduling response expectation constraint unit to obtain the nth behavior scheduling response expectation constraint; based on the behavior scheduling abnormality prediction and operation and maintenance unit, the target robot is subjected to abnormal prediction and operation and maintenance according to the nth skill behavior scheduling decision to generate the nth behavior scheduling abnormality prediction and operation and maintenance decision; the nth skill behavior scheduling decision, the nth behavior scheduling response expectation constraint and the nth behavior scheduling abnormality prediction and operation and maintenance decision are output as the nth task behavior scheduling strategy, and the nth task behavior scheduling strategy is added to the N task behavior scheduling strategies.

[0008] In a possible implementation, the nth skill behavior scheduling decision is subjected to scheduling response expectation constraints according to the behavior scheduling response expectation constraint unit to obtain the nth behavior scheduling response expectation constraint, and the following processing is also performed: the behavior scheduling response expectation constraint unit includes multiple behavior scheduling normal response record libraries corresponding to the multiple robots; based on the behavior scheduling response expectation constraint unit, the behavior scheduling normal response record library of the target robot is retrieved, wherein the behavior scheduling normal response record library of the target robot includes multiple behavior scheduling normal response records, wherein each behavior scheduling normal response record includes a skill behavior scheduling decision sample and a scheduling response normal monitoring sample; based on the nth skill behavior scheduling decision, a similarity analysis is performed on each skill behavior scheduling decision sample in the multiple behavior scheduling normal response records to obtain multiple behavior scheduling similarity coefficients; based on the multiple behavior scheduling similarity coefficients, the multiple behavior scheduling normal response records are monitored and sampled according to the behavior scheduling similarity constraints to establish a behavior scheduling response expectation space; a central trend analysis is performed according to the behavior scheduling response expectation space to generate the nth behavior scheduling response expectation constraint.

[0009] In a possible implementation, based on the behavior scheduling abnormality prediction and operation and maintenance unit, the target robot is subjected to abnormal prediction and operation and maintenance according to the nth skill behavior scheduling decision to generate the nth behavior scheduling abnormality prediction and operation and maintenance decision, and the following processing is also performed: the behavior scheduling abnormality prediction and operation and maintenance unit includes a scheduling abnormality simulation unit and a scheduling abnormality operation and maintenance map; based on the scheduling abnormality simulation unit, the target robot is subjected to abnormality simulation according to the nth skill behavior scheduling decision to obtain multiple scheduling abnormality simulation results; the multiple scheduling abnormality simulation results are input into the scheduling abnormality operation and maintenance map to obtain multiple scheduling abnormality prediction and operation and maintenance plans; the multiple scheduling abnormality simulation results and the multiple scheduling abnormality prediction and operation and maintenance plans are added to the nth behavior scheduling abnormality prediction and operation and maintenance decision.

[0010] In a possible implementation, the multiple robots are scheduled according to the application scenario task to determine the target robot, and the following processing is performed: the application scheduling bilateral constraints include application function association constraints and application time margin constraints; feature recognition is performed according to the application scenario task to obtain application task attributes and application task time limits; based on the application function association constraints, the multiple robots are scheduled according to the application task attributes to obtain a first scheduled robot distribution; based on the application time margin constraints, the multiple robots are scheduled according to the application task time limit to obtain a second scheduled robot distribution; the intersection of the first scheduled robot distribution and the second scheduled robot distribution is obtained to generate the target robot.

[0011] In a possible implementation, the multiple robots are scheduled according to the application task attributes to obtain a first scheduled robot distribution, and the following processing is performed: according to the SkiRobot platform, the functional characteristic parameters of the multiple robots are collected to obtain multiple robot functional characteristics; based on the application task attributes, the multiple robot functional characteristics are respectively associated and evaluated to obtain multiple application function association coefficients; it is determined whether the multiple application function association coefficients meet the application function association constraints to generate an application function association winning distribution; based on the application function association winning distribution, the multiple robots are screened to generate the first scheduled robot distribution.

[0012] In a possible implementation, based on the application time margin constraint, the multiple robots are scheduled according to the application task time limit to obtain a second scheduled robot distribution, and the following processing is also performed: according to the SkiRobot platform, the fixed task information of the multiple robots is collected to establish multiple fixed task tables; based on the multiple fixed task tables, the application time margin of the multiple robots is evaluated according to the application task time limit to obtain multiple application time margin coefficients; it is determined whether the multiple application time margin coefficients meet the application time margin constraint to generate an application time margin winning distribution; based on the application time margin winning distribution, the multiple robots are screened to generate the second scheduled robot distribution.

[0013] The present application also provides an electronic device, comprising: a memory for storing executable instructions; and a processor for implementing a robot behavior scheduling method combining modular skills when executing the executable instructions stored in the memory.

[0014] The present application also provides a computer-readable storage medium, including: a computer program stored thereon, which, when executed by a processor, implements a robot behavior scheduling method combining modular skills.

[0015] The robot behavior scheduling method, device and medium proposed in this application that combines modular skills are connected to the SkiRobot platform to receive task instructions; determine the target robot according to the application scenario task; disassemble the application scenario task to obtain N application subtasks; activate the robot behavior scheduling channel embedded in the SkiRobot platform; schedule each skill module of the target robot to establish a robot behavior scheduling tree; based on the SkiRobot platform, control the target robot to perform the application scenario task according to the robot behavior scheduling tree. This solves the technical problem that it is difficult to conveniently and flexibly realize complex behavior combinations and dynamic scheduling in the existing robot behavior scheduling, which leads to insufficient convenience, flexibility and low scene adaptability of robot behavior control, and achieves the technical effect of improving the convenience and flexibility of robot behavior control. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solution of the embodiment of the present disclosure, the accompanying drawings of the embodiment of the present disclosure will be briefly introduced below. A flow chart is used in the present application to illustrate the operations performed by the system according to the embodiment of the present application. It should be understood that the preceding or following operations are not necessarily performed accurately in order. On the contrary, various steps can be processed in reverse order or simultaneously as needed. At the same time, other operations can also be added to these processes, or one or more operations can be removed from these processes.

[0017] Figure 1A schematic diagram of the structure of a robot behavior scheduling method combined with modular skills provided in an embodiment of the present application;

[0018] Figure 2 A schematic diagram of a flow chart of establishing a robot behavior scheduling tree in a robot behavior scheduling method combined with modular skills provided in an embodiment of the present application;

[0019] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0020] Description of the reference numerals: input device 401 , processor 402 , memory 403 , output device 404 . DETAILED DESCRIPTION

[0021] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below.

[0022] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application will be further described in detail below in conjunction with the accompanying drawings. The described embodiments should not be regarded as limiting the present application. All other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of this application.

[0023] In the following description, reference is made to "some embodiments", which describe a subset of all possible embodiments, but it is understood that "some embodiments" may be the same subset or different subsets of all possible embodiments, and may be combined with each other without conflict, and the terms "first\second" involved are merely to distinguish similar objects and do not represent a specific ordering of objects. The terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products, or devices. Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those generally understood by technicians in the technical field of this application. The terms used herein are for the purpose of describing the embodiments of the present application only.

[0024] The present application embodiment provides a robot behavior scheduling method combining modular skills, such as Figure 1 As shown, including:

[0025] Step S100, connecting to the SkiRobot platform and receiving task instructions, wherein the task instructions include application scenario tasks. The SkiRobot platform is a robot control platform based on ROS2. The SkiRobot platform is communicatively connected to multiple control terminals of multiple robots, and each robot includes multiple skill modules.

[0026] Preferably, the task instructions are received and parsed by connecting to the SkiRobot robot control platform. Specifically, the SkiRobot platform is a robot control platform based on ROS2. The SkiRobot platform is also called the SkiROS2 platform. It is a skill-based ROS robot control platform. By modularizing skills and creating complex robot behaviors, it aims to provide flexible and modular skill scheduling and behavior management. Among them, ROS2 is the second-generation version of the robot operation control system. As a distributed middleware framework for developing robot applications, the SkiRobot platform is connected to the control terminals of multiple robots through communication protocols (such as the ROS2 messaging system). These terminals are system components of the robot and are used to receive instructions from the platform and control the operation of the robot. The task instruction is a set of received instructions, which includes specific tasks to be performed in a specific application scenario. It is usually sent to the SkiRobot platform by users or upper-level applications to instruct the robot to perform specific behaviors or tasks, such as moving objects, path planning, area cleaning, etc.

[0027] Preferably, the skill modules included in each robot are specific functional units that define the robot, and each skill module focuses on completing a specific, independent task. A robot may include a movement skill module, a grasping skill module, an object recognition skill module, etc., such as object recognition, path planning, moving to a location, grasping an object, area recognition, cleaning planning, navigating to an area, performing cleaning, etc. The skill modules enable the robot to have the ability to perform specific tasks and can be combined to complete complex tasks. For example, in a logistics warehouse robot project, the object recognition skill module can be reused in different warehouse environments and only needs to be fine-tuned according to the specific characteristics of the item; by combining multiple skill modules according to certain logical relationships, complex robot behaviors can be constructed, just like building blocks. According to different task requirements, skill modules can be flexibly selected and combined to quickly construct a robot behavior scheduling tree that meets the requirements. For example, the cleaning task of a home service robot can be composed of multiple skill modules such as area recognition, cleaning planning, navigating to an area, and performing cleaning.

[0028] Step S200, based on the application scheduling bilateral constraints, the multiple robots are scheduled according to the application scenario tasks to determine the target robot.

[0029] Preferably, when performing a specific task, a bilateral constraint method is used to select and schedule a robot suitable for task execution. Specifically, bilateral constraints refer to two-way constraints that consider both task requirements and robot capabilities during the scheduling process. That is, when selecting a robot, it is necessary to meet the requirements of the task (such as skills, time limits, execution conditions), and also to consider the availability, status, and current load of the robot, such as the skills required for the task, the power status of the robot, task priority, path accessibility, execution time window, etc. Application scenario tasks refer to specific tasks that need to be performed, which describe the goals to be achieved in a certain application scenario, such as cargo handling in a warehouse, item detection on a production line, etc. Each application scenario task has specific requirements and execution conditions. The scene task schedules multiple robots, that is, assigns tasks to multiple available robots and selects the most suitable executor based on task requirements and robot status. The application scheduling process ensures that task allocation is efficient and feasible by analyzing task requirements and robot characteristics, and finally determines the target robot that is most suitable for performing the current task among multiple robots. For example, in a warehouse automation system, if there is a task to move cargo A to a specified location, the application scheduling bilateral constraints will select the target robot based on the following factors: task constraints include the skills required for the task, such as path planning, object grasping and handling; robot constraints include the current status of the robot, such as whether it has the necessary skill modules (such as grasping arms and navigation capabilities), whether the battery is sufficient, whether it is currently idle, etc.

[0030] Furthermore, step S200 also includes step S210, wherein the application scheduling bilateral constraints include application function association constraints and application time margin constraints; step S220, performing feature recognition according to the application scenario tasks to obtain application task attributes and application task time limits; step S230, based on the application function association constraints, scheduling the multiple robots according to the application task attributes to obtain a first scheduling robot distribution; step S240, based on the application time margin constraints, scheduling the multiple robots according to the application task time limit to obtain a second scheduling robot distribution; step S250, obtaining the intersection of the first scheduling robot distribution and the second scheduling robot distribution to generate the target robot.

[0031] Preferably, the application scheduling bilateral constraints include application function association constraints and application time margin constraints, wherein the application function association constraints are used to schedule robots that meet these functional requirements according to task attributes (such as the skills and functional requirements required for the task, etc.), to ensure that the selected robots have the skills and capabilities required to complete the task; the application time margin constraints are used to schedule robots that can complete the task within the specified time according to the time requirements of the task (such as the deadline or response time of the task), to ensure that the robot can complete the task efficiently within the task time limit and avoid delays; specifically, the specific requirements and characteristics of the task are analyzed, the attributes of the task (such as the complexity of the task, the required skills, etc.) and the time limit (the time range or deadline within which the task must be completed) are obtained, and the application task attributes and the application task time limit, that is, the function of the task, are obtained. Functional attributes (such as whether handling, identification, path planning, etc. are required) and time constraints (such as the task must be completed within 30 minutes); functional association constraints are performed according to task attributes to screen out robots that meet these functional requirements, and the output is the distribution of robots that meet the functional requirements of the task. Then, check whether each robot can complete the task within the time specified in the task, and apply the time margin constraint to screen out robots that can complete the task within the time limit, and output the distribution of robots that meet the time constraints; then perform an intersection operation on the first scheduling robot distribution obtained according to the functional association constraint and the second scheduling robot distribution obtained according to the time margin constraint to find out the robot that meets both the task functional requirements and the time requirements, and finally generate a target robot that meets both the application function and time constraints for performing the task.

[0032] Furthermore, step S230 also includes step S231, collecting functional characteristic parameters of the multiple robots according to the SkiRobot platform to obtain multiple robot functional characteristics; step S232, based on the application task attributes, respectively performing correlation evaluation on the multiple robot functional characteristics to obtain multiple application function correlation coefficients; step S233, judging whether the multiple application function correlation coefficients meet the application function correlation constraints, and generating an application function correlation winning distribution; step S234, based on the application function correlation winning distribution, screening the multiple robots to generate the first scheduling robot distribution.

[0033] Preferably, the functional characteristic data of multiple robots are obtained using the SkiRobot platform, which may include the skill module, execution capability, sensor configuration, load capacity, speed and accuracy of each robot, and the functional characteristics of multiple robots are obtained. The task attributes (such as specific functions and skill modules required for the task) are analyzed and applied, and they are associated with the functional characteristics of each robot for evaluation, and multiple application function association coefficients are obtained, which represent the matching degree between each robot and the functional requirements of the application task. Each association coefficient measures the suitability of the robot in meeting the task function attributes, and the higher the value, the better the matching degree; then the application function association coefficient of each robot is compared with the preset application function association constraint conditions to determine which robots meet these conditions, wherein the application function association constraint defines the minimum functional matching standard that the robot must meet, and the robots that meet the application function association constraint are marked as winning robots, forming a winning distribution, showing which robots perform best in terms of functional matching; finally, robots that meet the task requirements are selected from the winning distribution, and those robots that do not meet the requirements are eliminated. The final set of robots selected forms the first scheduling robot distribution, indicating that these robots are most suitable for performing application tasks in terms of function.

[0034] Furthermore, step S240 also includes step S241, collecting fixed task information of the multiple robots according to the SkiRobot platform, and establishing multiple fixed task tables; step S242, based on the multiple fixed task tables, evaluating the application time margin of the multiple robots according to the application task time limit, and obtaining multiple application time margin coefficients; step S243, judging whether the multiple application time margin coefficients meet the application time margin constraints, and generating an application time margin winning distribution; step S244, based on the application time margin winning distribution, screening the multiple robots, and generating the second scheduling robot distribution.

[0035] Preferably, the SkiRobot platform collects fixed task information of each robot, i.e., assigned task information, including historical data of the robot when performing similar tasks, such as the time required to complete the task, the average response time of task execution, the task queue, etc., and establishes multiple fixed task tables based on the collected information, each table containing the time and related data required for the robot to complete different tasks, and then uses the application task time limit to analyze the fixed task tables of each robot to evaluate whether they can complete the task within a given time, and obtain multiple application time margin coefficients, which represent the adaptability of each robot in meeting the task time limit. The higher the coefficient, the better the performance. The greater the possibility that the robot can complete the task within the task time limit; then compare the application time margin coefficient of each robot with the preset application time margin constraint to determine which robots can complete the task within the specified time, where the time margin constraint sets the time requirement for task completion, such as the task must be completed within 30 minutes. According to the comparison results, a winning distribution of application time margin is generated, and the robots that meet the time constraint are listed. Finally, the robots that meet the time constraint are screened out according to the winning distribution, and the robots that do not meet the conditions are excluded to form the second scheduling robot distribution, which represents the robots that are most suitable for performing tasks in terms of time. For example, in warehouse management, if the task is "complete cargo handling within 15 minutes", the SkiRobot platform will collect the robot's historical task information (such as the time record of completing similar handling tasks in the past) and establish a fixed task table, evaluate the time margin of each robot based on the task time limit, calculate the application time margin coefficient, screen out the robots that are capable of completing the task within 15 minutes, and generate the second scheduling robot distribution.

[0036] Step S300, disassemble the application scenario task to obtain N application subtasks, where N is a positive integer greater than 1.

[0037] Preferably, a complex application scenario task is decomposed into multiple subtasks that are easy to manage so as to be executed more efficiently in the robot. Specifically, a complex application scenario task is subdivided into multiple subtasks to make each subtask simpler, clearer, and easier to assign and execute. For example, a warehouse handling task can be decomposed into the following subtasks: identifying the location of the target goods, planning a path from the current robot position to the goods location, moving to the target goods location, performing the operation of grabbing the goods, planning a path from the goods location to the target placement location, moving to the target placement location and placing the goods. N represents the number of subtasks obtained after the task is decomposed, which is a positive integer greater than 1, indicating that the application scenario task is decomposed into at least two steps, each step is an independent subtask, and these subtasks can be assigned to the same robot for sequential execution, or assigned to multiple robots to complete in parallel to speed up the overall completion of the task.

[0038] Step S400, activating the robot behavior scheduling channel embedded in the SkiRobot platform, wherein the robot behavior scheduling channel includes a skill behavior scheduling decision unit, a behavior scheduling response expectation constraint unit and a behavior scheduling abnormality prediction operation and maintenance unit.

[0039] Preferably, the robot behavior scheduling channel embedded in the SkiRobot platform is activated to ensure that the entire scheduling process from task reception to execution is optimally managed. The robot behavior scheduling channel is a core module in the SkiRobot platform, which is used to manage and coordinate the scheduling of robot behaviors. The robot behavior scheduling channel is responsible for receiving task instructions, decomposing and distributing tasks to appropriate robot skill modules for execution, and through the coordination of various scheduling units, making robot behavior scheduling more efficient and reliable. The robot behavior scheduling channel includes a skill behavior scheduling decision unit, a behavior scheduling response expectation constraint unit, and a behavior scheduling anomaly prediction and operation unit. Specifically, the skill behavior scheduling decision unit is a key component in the behavior scheduling channel, which is responsible for formulating and adjusting the scheduling decisions of the robot skill modules based on the task instructions and the current system status. The unit will evaluate the task requirements and the current capabilities of the robot, select the most appropriate skill module for task execution, and ensure that the robot performs the task under the optimal path and conditions according to the requirements of the task and scenario, including selecting the task sequence, optimizing the execution steps, and determining the scheduling strategy of the task.

[0040] Preferably, the behavior scheduling response expectation constraint unit is used to define and manage the expected behavior and response of the robot when performing a task, and includes a normal response record library for behavior scheduling, which constrains and detects the normality of behavior scheduling based on historical data and preset standards, that is, during the task execution process, by analyzing historical scheduling records and current behaviors, it ensures that the robot behavior meets expectations, so as to improve the consistency and accuracy of task execution and ensure that the task does not deviate from the normal behavior trajectory; the behavior scheduling anomaly prediction and operation and maintenance unit is the component in the scheduling channel responsible for predicting and handling potential scheduling problems. It identifies and predicts possible anomalies or failures by analyzing data in the scheduling process. Its functions include monitoring the behavior scheduling process, identifying scheduling anomalies in advance, such as robot task failure or inconsistent response, and taking preventive or corrective measures when necessary to reduce the risk of task execution interruption or failure.

[0041] Step S500, based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, schedule each skill module of the target robot according to the N application subtasks to establish a robot behavior scheduling tree.

[0042] Preferably, the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling anomaly prediction operation and maintenance unit are used to schedule the various skill modules of the target robot according to N application subtasks, that is, according to the N application subtasks, determine which skill modules need to be activated, how to arrange their execution order, and how to coordinate the operation of different modules during task execution, and establish a robot behavior scheduling tree, that is, the logical structure of the task is represented by a tree combination, the root node of the scheduling tree represents the entire application scenario task, each branch node represents a different application subtask, and the leaf node corresponds to a specific skill module execution. Specifically, the various scheduling units work together, such as the skill behavior scheduling decision unit assigns tasks, the behavior scheduling response expectation constraint unit ensures the execution standard, and the behavior scheduling anomaly prediction operation and maintenance unit monitors the stability of the execution process. Based on the logical relationship between the subtasks, a topological connection is constructed to form a complete behavior scheduling tree. The connection relationship in the tree represents the execution order and conditions of the tasks.

[0043] Preferably, a plurality of skill modules are combined according to certain logical relationships through a behavior scheduling tree to construct a complex robot behavior. The root node of the behavior scheduling tree is the overall task goal of the robot, the leaf nodes are specific skill modules, and the intermediate nodes represent logical relationships and decision points. For example, in a warehouse item handling task, the root node of the behavior scheduling tree can be connected in sequence to skill modules such as item recognition, path planning, moving position, and grabbing items. The robot executes these skills in sequence according to the structure of the behavior scheduling tree to complete the item handling task. The skill modules can work together through data transmission and sharing, and the output result of one skill module can be used as the input of another skill module, thereby realizing the flow of information and the connection of tasks. For example, after the item recognition skill module recognizes the location information of the target item, it passes this information to the path planning skill module, and the path planning module plans the robot's moving path based on this location information.

[0044] Preferably, the behavior scheduling tree is a hierarchical structure that intuitively expresses the behavior logic of the robot in different situations, can clearly see the decision-making process and execution order of the robot's behavior, and is easy to understand and debug. For example, under the root node of the behavior scheduling tree, different sub-nodes can be selected for execution according to different conditional branches, thereby realizing the robot's intelligent decision-making. The root node can determine whether it is necessary to perform the item handling task based on environmental information. If necessary, it executes the item recognition, path planning, moving position, grabbing items and other sub-nodes in sequence; the structure of the behavior scheduling tree enables the robot to quickly switch behaviors according to real-time environmental changes. When a sub-node fails to execute, the behavior scheduling tree can automatically switch to other sub-nodes according to the preset logic to continue execution, thereby ensuring that the robot's behavior can adapt to different situations. For example, if the item recognition sub-node fails to successfully identify the item when the robot is performing the item handling task, the behavior scheduling tree can automatically switch to the sub-node of re-identifying the item or finding other items, thereby improving the robot's adaptability and reliability.

[0045] Further, such as Figure 2 As shown, step S500 also includes step S510, based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, the respective skill modules are scheduled according to the N application subtasks to obtain N task behavior scheduling strategies; step S520, setting the application scenario task as the scheduling tree root node; step S530, taking the N application subtasks as N scheduling branch nodes, and taking the N task behavior scheduling strategies as N scheduling leaf nodes; step S540, taking the logical relationship between the application scenario task, the N application subtasks and the N task behavior scheduling strategies as the scheduling tree topological relationship; step S550, based on the scheduling tree topological relationship, topologically connecting the scheduling tree root node, the N scheduling branch nodes and the N scheduling leaf nodes to generate the robot behavior scheduling tree.

[0046] Preferably, the application scenario task is decomposed into N subtasks, and a suitable skill module (skill behavior scheduling decision unit, behavior scheduling response expectation constraint unit and behavior scheduling abnormality prediction operation and maintenance unit) is specified for each subtask for scheduling, and a task behavior scheduling strategy is generated for each subtask, that is, N task behavior scheduling strategies are obtained, and the application scenario task is used as the root node of the scheduling tree, representing the execution goal of the entire task, and then the N application subtasks and the corresponding N task behavior scheduling strategies are used as branch nodes and leaf nodes of the tree structure respectively, where the branch node represents the decomposition structure of the task, each node corresponds to a subtask, and the leaf node represents The specific execution strategy of each subtask and the arrangement of skill modules; the logical relationship between application scenario tasks, subtasks, and task behavior scheduling strategies defines the topological structure of the behavior scheduling tree, showing how the task is decomposed from the overall to the specific execution strategy, and completed through the scheduling of skill modules, to obtain the scheduling tree topological relationship, that is, the connection relationship between the root node, branch node and leaf node, explaining the order and conditions of task execution, and constructing a complete robot behavior scheduling tree by connecting the root node, subtask node and leaf node according to the logical relationship, describing the execution details from the start of the task to each subtask, and then to the specific skill module execution strategy.

[0047] Furthermore, step S510 also includes step S511, extracting the nth application subtask according to the N application subtasks, wherein n is a positive integer, 1≤n≤N; step S512, performing association analysis on the various skill modules based on the nth application subtask to determine the nth task-associated skill module; step S513, collecting basic information of the nth task-associated skill module according to the SkiRobot platform to obtain the nth skill module data; step S514, inputting the nth application subtask and the nth skill module data into the skill behavior scheduling decision unit to obtain the nth skill behavior scheduling decision; step S515, based on According to the behavior scheduling response expectation constraint unit, the nth skill behavior scheduling decision is subjected to the scheduling response expectation constraint to obtain the nth behavior scheduling response expectation constraint; step S516, based on the behavior scheduling abnormality prediction and operation and maintenance unit, according to the nth skill behavior scheduling decision, the target robot is subjected to abnormal prediction and operation and maintenance to generate the nth behavior scheduling abnormality prediction and operation and maintenance decision; step S517, the nth skill behavior scheduling decision, the nth behavior scheduling response expectation constraint and the nth behavior scheduling abnormality prediction and operation and maintenance decision are output as the nth task behavior scheduling strategy, and the nth task behavior scheduling strategy is added to the N task behavior scheduling strategies.

[0048] Preferably, the nth subtask is randomly extracted from the N application subtasks for analysis and processing, that is, the nth application subtask is analyzed to identify the skill module associated with it. For example, if the nth subtask is "identifying objects", the skill modules that need to be used, such as camera modules, image processing modules, etc., are analyzed, and through association analysis, it is determined which skill modules need to be activated and scheduled when executing the subtask; then the basic information of these associated skill modules, including the status, capability parameters, current load, etc. of the skill modules are collected and analyzed through the SkiRobot platform, and the corresponding nth skill module data is generated, and then the nth application subtask is assigned to the SkiRobot platform. The subtask and the nth skill module data are input together into the skill behavior scheduling decision unit, and the scheduling decision for the nth subtask is evaluated and output, such as starting camera recognition and saving the result. A decision model is built based on a neural network model such as a recursive neural network (RNN) or a long short-term memory network (LSTM), and input data (such as task description, skill module status, sensor data, etc.) and output data (such as skill scheduling strategy) are obtained. Preprocessing is performed separately to ensure data quality, and relevant features are extracted, such as task type, resource status, environmental conditions, etc. The decision model is trained to obtain a skill behavior scheduling decision unit.

[0049] Preferably, the generated nth skill behavior scheduling decision is input into the behavior scheduling response expectation constraint unit, which performs analysis based on historical data and expected standards, sets the response expectation of the task, ensures that the execution meets expectations, and outputs a set of standards and conditions for evaluating and verifying the task execution results, i.e., obtaining the nth behavior scheduling response expectation constraint; then, the nth skill behavior scheduling decision is analyzed using the behavior scheduling abnormality prediction operation and maintenance unit to predict possible abnormal situations, such as task delays, equipment failures, etc., thereby generating the nth behavior scheduling abnormality prediction operation and maintenance decision, i.e., formulating early warning and emergency measures to prevent and respond to potential execution problems; the generated nth skill behavior scheduling decision, the nth behavior scheduling response expectation constraint, and the nth behavior scheduling abnormality prediction operation and maintenance decision are integrated into a complete nth task behavior scheduling strategy, and the currently generated nth task behavior scheduling strategy is added to the behavior scheduling strategy set of the overall task, so as to gradually build a complete scheduling plan to ensure efficient execution and stability of the task.

[0050] Further, step S515 also includes step S5151, wherein the behavior scheduling response expectation constraint unit includes multiple behavior scheduling normal response record libraries corresponding to the multiple robots; step S5152, based on the behavior scheduling response expectation constraint unit, calling the behavior scheduling normal response record library of the target robot, wherein the behavior scheduling normal response record library of the target robot includes multiple behavior scheduling normal response records, wherein each behavior scheduling normal response record includes a skill behavior scheduling decision sample and a scheduling response normal monitoring sample; step S5153, based on the nth skill behavior scheduling decision, performing similarity analysis on each skill behavior scheduling decision sample in the multiple behavior scheduling normal response records to obtain multiple behavior scheduling similarity coefficients; step S5154, based on the multiple behavior scheduling similarity coefficients, screening monitoring samples of the multiple behavior scheduling normal response records according to the behavior scheduling similarity constraints, and establishing a behavior scheduling response expectation space; step S5155, performing a central trend analysis based on the behavior scheduling response expectation space to generate the nth behavior scheduling response expectation constraint.

[0051] Preferably, the behavior scheduling response expectation constraint unit is composed of multiple behavior scheduling normal response record libraries corresponding to multiple robots, each record library contains historical behavior data related to a specific robot, which is used to analyze and verify whether the execution of the robot task meets expectations, and the behavior scheduling normal response record library related to the target robot is called out from the behavior scheduling response expectation constraint unit. This record library stores multiple behavior scheduling normal response records of the robot in the execution of historical tasks. Each behavior scheduling normal response record contains two parts, a skill behavior scheduling decision sample and a scheduling response normal monitoring sample, wherein the skill behavior scheduling decision sample describes the specific skill scheduling strategy used in the historical task, and the scheduling response normal monitoring sample reflects the actual response during the task execution, showing how the task is successfully executed.

[0052] Preferably, a similarity analysis is performed between the current nth skill behavior scheduling decision and each skill behavior scheduling decision sample in the target robot history record library, and the similarity is calculated to quantify the similarity between the current scheduling decision and the historical decision, and a set of behavior scheduling similarity coefficients is generated. The higher the similarity, the closer the current task is to the historical task. Then, based on multiple behavior scheduling similarity coefficients and applying behavior scheduling similarity constraints, the historical monitoring samples closest to the current task are screened out, and a behavior scheduling response expectation space is established, that is, the scheduling response normal monitoring samples corresponding to the behavior scheduling similarity coefficients that meet the behavior scheduling similarity constraints are added to the behavior scheduling response expectation space, indicating the expected response range of the task in similar historical tasks. A central trend analysis is performed on the data in the behavior scheduling response expectation space, such as calculating the mean, median or mode, to summarize the main characteristics of the space, and based on the analysis results, the response expectation constraint for the nth skill behavior scheduling is generated.

[0053] Furthermore, step S516 also includes step S5161, wherein the behavior scheduling abnormality prediction and operation and maintenance unit includes a scheduling abnormality simulation unit and a scheduling abnormality operation and maintenance map; step S5162, based on the scheduling abnormality simulation unit, the target robot is subjected to an abnormality simulation according to the nth skill behavior scheduling decision to obtain a plurality of scheduling abnormality simulation results; step S5163, the plurality of scheduling abnormality simulation results are input into the scheduling abnormality operation and maintenance map to obtain a plurality of scheduling abnormality prediction and operation and maintenance plans; step S5164, the plurality of scheduling abnormality simulation results and the plurality of scheduling abnormality prediction and operation and maintenance plans are added to the nth behavior scheduling abnormality prediction and operation and maintenance decision.

[0054] Preferably, the scheduling exception simulation unit and the scheduling exception operation and maintenance map constitute a behavior scheduling exception prediction operation and maintenance unit. The scheduling exception simulation unit is used to simulate abnormal situations that may occur in robot tasks to identify potential risks and problems. The scheduling exception operation and maintenance map contains exception handling solutions, solution strategies and related data in historical tasks, which are used to guide and generate solutions to deal with abnormalities. Specifically, when performing task scheduling, based on the current n-th skill behavior scheduling decision, an abnormality simulation is performed through the scheduling exception simulation unit to analyze the problems that may occur during the execution of the task, such as hardware failure, task interruption or external environment interference, and generate multiple scheduling exception simulation results. The output results of the abnormal simulation are then input into the scheduling exception operation and maintenance map for matching and analysis to generate multiple scheduling exception prediction operation and maintenance solutions, including specific steps and strategies for how to deal with different abnormal situations that occur in the simulation. Finally, the multiple scheduling exception simulation results and the multiple scheduling exception prediction operation and maintenance solutions are added to the n-th behavior scheduling exception prediction operation and maintenance decision, covering the abnormalities that may be encountered in the execution of the task and their corresponding countermeasures, which are used for real-time adjustment and optimization during the actual execution of the task, so as to quickly respond to and handle emergencies and maintain the stability and reliability of the task.

[0055] Step S600: Based on the SkiRobot platform, the target robot is controlled to execute the application scenario task according to the robot behavior scheduling tree.

[0056] Preferably, the SkiRobot platform controls the robot to complete each step of the task in sequence through the logical structure of the scheduling tree, wherein the scheduling tree instructs the robot to start from the root node and execute each subtask and skill module in sequence or conditionally. Specifically, during the task execution process, the SkiRobot platform monitors the robot's behavior and execution in real time. Based on the paths and nodes in the scheduling tree, the platform can judge the progress of the current task and make necessary adjustments based on the feedback information. If an exception occurs when executing a certain node, the scheduling tree combined with the behavior scheduling exception prediction operation and maintenance unit can guide the robot to switch to an alternative path or retry execution. , thereby maintaining the continuity and stability of the task; after the platform receives the task instruction, it activates the corresponding behavior scheduling tree and sends instructions to the target robot to start execution. According to the structure of the scheduling tree, the robot performs subtasks in sequence, for example, starting from the "identify object" node, and automatically transfers to the "path planning" node after completion, and then to the "move to the target location", etc. After each node is executed, the platform checks whether it meets the expected response (monitored by the behavior scheduling response expectation constraint unit). If an abnormality or failure is detected, the abnormal prediction operation and maintenance unit is used to take countermeasures, such as retrying or selecting an alternative path, and finally completing the application scenario task efficiently and structuredly.

[0057] Based on the foregoing embodiments, the embodiments of the present application further provide an electronic device and a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor of the electronic device, it can implement the method described in any of the foregoing embodiments.

[0058] Figure 3 1 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, showing a block diagram of an exemplary electronic device suitable for implementing an implementation of the present invention. Figure 3 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention. The electronic device is in the form of a general computing device, and its components may include but are not limited to an input device 401, a processor 402, a memory 403, and an output device 404. Among them, the processor 402 may be one or more; the memory 403 may include a computer-readable medium and at least one program product, and the program product has a set (at least one) of program modules, which are configured to perform the functions of each embodiment of the present application.

[0059] The memory 403 shown in the embodiment of the present invention may adopt any combination of one or more computer-readable media; the computer-readable storage medium may be, but is not limited to, an infrared, semiconductor system, device or component, or any combination of the above, for storing software programs, computer executable programs and modules, such as the program instructions / modules corresponding to a robot behavior scheduling method combined with modular skills in an embodiment of the present invention. The processor 402 executes various functional applications and data processing of the computer device by running the software programs, instructions and modules stored in the memory 403, thereby realizing the above-mentioned robot behavior scheduling method combined with modular skills.

[0060] Although the present application makes various references to certain modules in the embodiments of the present application, any number of different modules may be used and run on a user terminal and / or server, and the various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of the functional units are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of the present invention.

[0061] The above specific implementations do not constitute a limitation on the protection scope of this application. It should be understood by those skilled in the art that various modifications, combinations and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions and improvements made within the spirit and principles of this application should be included in the protection scope of this application.

Claims

1. A robot behavior scheduling method combining modular skills, characterized in that: The method comprises: Connecting to the SkiRobot platform to receive task instructions, wherein the task instructions include application scenario tasks, the SkiRobot platform is a robot control platform based on ROS2, the SkiRobot platform is communicatively connected with multiple control terminals of multiple robots, and each robot includes multiple skill modules; Based on the application scheduling bilateral constraints, the multiple robots are scheduled according to the application scenario tasks to determine the target robot, wherein the application scheduling bilateral constraints include application function association constraints and application time margin constraints; Decomposing the application scenario task to obtain N application subtasks, where N is a positive integer greater than 1; Activate the robot behavior scheduling channel embedded in the SkiRobot platform, wherein the robot behavior scheduling channel includes a skill behavior scheduling decision unit, a behavior scheduling response expectation constraint unit, and a behavior scheduling abnormality prediction operation and maintenance unit; Based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, each skill module of the target robot is scheduled according to the N application subtasks to establish a robot behavior scheduling tree; Based on the SkiRobot platform, the target robot is controlled to execute the application scenario task according to the robot behavior scheduling tree.

2. A robot behavior scheduling method combining modular skills as claimed in claim 1, characterized in that: Based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, each skill module of the target robot is scheduled according to the N application subtasks to establish a robot behavior scheduling tree, including: Based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, each skill module is scheduled according to the N application subtasks to obtain N task behavior scheduling strategies; Setting the application scenario task as the scheduling tree root node; The N application subtasks are N scheduling tree branch nodes, and the N task behavior scheduling strategies are N scheduling tree leaf nodes; The logical relationship between the application scenario task, the N application subtasks and the N task behavior scheduling strategies is used as a scheduling tree topology relationship; Based on the scheduling tree topological relationship, the scheduling tree root node, the N scheduling tree branch nodes and the N scheduling tree leaf nodes are topologically connected to generate the robot behavior scheduling tree.

3. A robot behavior scheduling method combining modular skills as claimed in claim 2, characterized in that: Based on the skill behavior scheduling decision unit, the behavior scheduling response expectation constraint unit and the behavior scheduling abnormality prediction operation and maintenance unit, the skill modules are scheduled respectively according to the N application subtasks to obtain N task behavior scheduling strategies, including: Extracting an nth application subtask according to the N application subtasks, where n is a positive integer, 1≤n≤N; Performing association analysis on the skill modules based on the nth application subtask to determine the skill module associated with the nth task; According to the SkiRobot platform, basic information of the skill module associated with the nth task is collected to obtain the nth skill module data; Inputting the nth application subtask and the nth skill module data into the skill behavior scheduling decision unit to obtain the nth skill behavior scheduling decision; Performing a scheduling response expectation constraint on the nth skill behavior scheduling decision according to the behavior scheduling response expectation constraint unit to obtain the nth behavior scheduling response expectation constraint; Based on the behavior scheduling abnormality prediction and operation and maintenance unit, performing abnormality prediction and operation and maintenance on the target robot according to the nth skill behavior scheduling decision, and generating an nth behavior scheduling abnormality prediction and operation and maintenance decision; The nth skill behavior scheduling decision, the nth behavior scheduling response expectation constraint and the nth behavior scheduling abnormality prediction operation and maintenance decision are output as the nth task behavior scheduling strategy, and the nth task behavior scheduling strategy is added to the N task behavior scheduling strategies.

4. A robot behavior scheduling method combining modular skills as claimed in claim 3, characterized in that: Performing a scheduling response expectation constraint on the nth skill behavior scheduling decision according to the behavior scheduling response expectation constraint unit to obtain the nth behavior scheduling response expectation constraint includes: The behavior scheduling response expectation constraint unit includes a plurality of behavior scheduling normal response record libraries corresponding to the plurality of robots; Based on the behavior scheduling response expectation constraint unit, calling the behavior scheduling normal response record library of the target robot, wherein the behavior scheduling normal response record library of the target robot includes multiple behavior scheduling normal response records, wherein each behavior scheduling normal response record includes a skill behavior scheduling decision sample and a scheduling response normal monitoring sample; Based on the nth skill behavior scheduling decision, a similarity analysis is performed on each skill behavior scheduling decision sample in the plurality of behavior scheduling normal response records to obtain a plurality of behavior scheduling similarity coefficients; Based on the multiple behavior scheduling similarity coefficients, the multiple behavior scheduling normal response records are screened for monitoring samples according to the behavior scheduling similarity constraints to establish a behavior scheduling response expectation space; A central tendency analysis is performed according to the behavior scheduling response expectation space to generate the nth behavior scheduling response expectation constraint.

5. A robot behavior scheduling method combining modular skills as claimed in claim 3, characterized in that: Based on the behavior scheduling abnormality prediction operation and maintenance unit, performing abnormality prediction operation and maintenance on the target robot according to the nth skill behavior scheduling decision, generating the nth behavior scheduling abnormality prediction operation and maintenance decision, including: The behavior scheduling anomaly prediction operation and maintenance unit includes a scheduling anomaly simulation unit and a scheduling anomaly operation and maintenance map; Based on the scheduling abnormality simulation unit, performing abnormality simulation on the target robot according to the nth skill behavior scheduling decision to obtain multiple scheduling abnormality simulation results; Inputting the plurality of scheduling anomaly simulation results into the scheduling anomaly operation and maintenance map to obtain a plurality of scheduling anomaly prediction operation and maintenance solutions; The multiple scheduling anomaly simulation results and the multiple scheduling anomaly prediction and operation and maintenance solutions are added to the nth line as a scheduling anomaly prediction and operation and maintenance decision.

6. A robot behavior scheduling method combining modular skills as claimed in claim 1, characterized in that: Based on the application scheduling bilateral constraints, the multiple robots are scheduled according to the application scenario tasks to determine the target robot, including: Perform feature recognition according to the application scenario task to obtain application task attributes and application task time limit; Based on the application function association constraint, scheduling the multiple robots according to the application task attributes to obtain a first scheduling robot distribution; Based on the application time margin constraint, scheduling the multiple robots according to the application task time limit to obtain a second scheduled robot distribution; The intersection of the first dispatching robot distribution and the second dispatching robot distribution is obtained to generate the target robot.

7. A robot behavior scheduling method combining modular skills as claimed in claim 6, characterized in that: Based on the application function association constraint, the plurality of robots are scheduled according to the application task attribute to obtain a first scheduled robot distribution, including: According to the SkiRobot platform, collecting functional characteristic parameters of the multiple robots to obtain multiple robot functional characteristics; Based on the application task attributes, respectively, the plurality of robot functional characteristics are evaluated for association to obtain a plurality of application function association coefficients; Determining whether the plurality of application function association coefficients satisfy the application function association constraint, and generating an application function association winning distribution; Based on the application function associated winning distribution, the multiple robots are screened to generate the first scheduling robot distribution.

8. A robot behavior scheduling method combining modular skills as claimed in claim 6, characterized in that: Based on the application time margin constraint, the plurality of robots are scheduled according to the application task time limit to obtain a second scheduled robot distribution, including: According to the SkiRobot platform, collecting fixed task information of the multiple robots and establishing multiple fixed task tables; Based on the multiple fixed task tables, and according to the application task time limit, the multiple robots are evaluated for application time margin to obtain multiple application time margin coefficients; Determining whether the multiple application time margin coefficients satisfy the application time margin constraint, and generating an application time margin superior distribution; Based on the application time margin superior distribution, the multiple robots are screened to generate the second scheduling robot distribution.

9. An electronic device, characterized in that: The electronic device comprises: A memory for storing executable instructions; The processor is used to implement a robot behavior scheduling method combining modular skills as described in any one of claims 1 to 8 when executing executable instructions stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, a robot behavior scheduling method combining modular skills is implemented as described in any one of claims 1-8.

Images