Multi-robot synchronous control method and system

Abstract

The invention discloses a multi-robot synchronous control method and system, and relates to the technical field of robot control, and the method comprises the steps: uploading a directional task to a task management center, carrying out the task interpretation and role marshalling planning through a task engine, determining an execution unit cluster, and building a task communication system based on the execution unit cluster through a 5G private network. Gathering at a task side end and deploying a dynamic management component; and based on the task communication system and the dynamic management component, the execution unit cluster is taken over, and driving control and tracking supervision are carried out by robot coding, so that the technical problems of relatively limited efficiency, flexibility and expandability of robot cluster task control and scheduling in a complex environment in the prior art are solved. And efficient, flexible and extensible task management and scheduling can be realized in a complex environment.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot control, in particular to a multi-robot synchronous control method and system. BACKGROUND

[0002] In the current field of multi-robot cooperation, especially in the scenarios of industrial automation and complex task execution, task allocation and synchronous control still face many challenges. Most existing technical solutions rely on single robots to execute tasks or cooperate in small-scale robot groups, which cannot effectively cope with the complexity and dynamic changes in multi-task parallelism and large-scale robot cooperation.

[0003] At the same time, traditional multi-robot systems often rely on manual intervention or simple preset rules for task allocation, resulting in low efficiency, poor flexibility and difficulty in handling unexpected situations, lacking real-time feedback and dynamic adjustment capabilities for environmental changes and robot states. In addition, during the execution process, the differences in the functions and capabilities of each robot make it difficult to reasonably balance the task coordination. In summary, the existing technology cannot flexibly cope with complex environmental changes and task requirements in multi-robot control, and there is still a large technical gap in dynamic adjustment of task execution, flexible transformation of robot roles, and rapid response to tasks.

[0004] Therefore, there is an urgent need for a multi-robot cluster-based synchronous control method that can achieve efficient, flexible and scalable task management and scheduling in complex environments. SUMMARY

[0005] The present application provides a multi-robot synchronous control method and system to solve the technical problem of limited efficiency, flexibility and scalability of robot cluster task control and scheduling in complex environments in the prior art.

[0006] In view of the above problems, the present application provides a multi-robot synchronous control method and system.

[0007] In a first aspect, the present application provides a multi-robot synchronous control method, which comprises uploading a directional task to a task management center, interpreting the task with a task engine and planning role grouping, determining an execution unit cluster, wherein the role grouping is performed in a function-oriented master-slave cooperation mode, and the robot code is used as the execution unit; a task communication system based on the execution unit cluster is established through a 5G private network, and a dynamic management component is deployed at the task edge and gathered; based on the task communication system and the dynamic management component, the execution unit cluster is taken over, and the robot code is used for driving control and tracking supervision, wherein the role grouping is re-distributed and reconstructed under abnormal working conditions, and the safety isolation control is performed.

[0008] On the second aspect, the present application provides a multi-robot synchronous control system, which includes: a task planning unit for uploading directed tasks to a task management center, using a task engine to interpret tasks and plan role groupings, and determining an execution unit cluster, wherein role grouping is performed based on function-oriented master-slave collaboration, and robot coding is used as the execution unit; a communication deployment unit for establishing a task communication system based on the execution unit cluster using a 5G private network, converging at the task edge and deploying dynamic management components; a control and supervision unit for taking over the execution unit cluster based on the task communication system and the dynamic management components, and using robot coding for drive control and tracking supervision, wherein robot task parameter control adjustment and grouping reconstruction and redistribution under fault blocking and safety isolation management are performed under abnormal working conditions.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] An embodiment of the present application provides a multi-robot synchronous control method, which uploads directed tasks to a task management center, uses a task engine to interpret tasks and plan role grouping, determines an execution unit cluster, and uses a 5G private network to establish a task communication system based on the execution unit cluster, converges at the task edge and deploys a dynamic management component; based on the task communication system and the dynamic management component, takes over the execution unit cluster, and uses robot coding for drive control and tracking supervision, which is used to solve the technical problems of the existing technology in the efficiency, flexibility and scalability of robot cluster task control and scheduling in complex environments, and can achieve efficient, flexible and scalable task management and scheduling in complex environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 A schematic flow chart of a multi-robot synchronous control method is provided for this application;

[0012] Figure 2 A schematic structural diagram of a multi-robot synchronous control system is provided for this application.

[0013] Description of the reference numerals: mission planning unit 11 , communication deployment unit 12 , control and supervision unit 13 . DETAILED DESCRIPTION

[0014] This application provides a multi-robot synchronous control method and system to solve the technical problems in the prior art of limited efficiency, flexibility and scalability of robot cluster task control and scheduling in complex environments.

[0015] Example 1: Figure 1 As shown, the present application provides a multi-robot synchronous control method, the method comprising:

[0016] S1: uploading a directional task to a task management center for task interpretation and role grouping planning by a task engine, determining a cluster of execution units, wherein role grouping is performed in a function-oriented master-slave collaboration, and a robot code is used as an execution unit.

[0017] In this step, the directional task needs to be uploaded to the task management center first. The directional task refers to a task target that needs to be executed by multiple robots in collaboration, and its content covers specific requirements of the task, execution time, and resource requirements, etc. The task management center receives the task and provides a basis for subsequent task interpretation and allocation.

[0018] Next, the task management center interprets and plans the role grouping of the uploaded task through the task engine. The task management center is a micro-platform for robot clusters as task executors.

[0019] Specifically, task interpretation refers to converting the information in the directional task into a format that can be understood and processed, and determining the corresponding execution steps according to the task requirements. During task interpretation, the task engine needs to analyze the specific content, required functions, execution conditions, etc. of the task, and adapt the task to ensure smooth connection with the operation process of the robot.

[0020] Among them, role grouping planning refers to matching each functional module involved in the task with the robot population, and planning the role and responsibility of each robot in the task.

[0021] At this time, the task engine performs role grouping in a master-slave collaboration based on the principle of function orientation, i.e. the robots executing the task are divided into master robots and slave robots. The master robots are responsible for executing the main line of the task, while the slave robots execute auxiliary work according to the task rhythm of the master robots.

[0022] Preferably, to ensure the collaborative work of each execution unit, i.e. each robot in the task, the task engine assigns each execution unit based on a robot code as a unique identifier.

[0023] Among them, the robot code is used as a unique identifier of each robot in this scheme, and the task engine accurately determines the task content and execution order of each robot through this code. The use of robot code helps to achieve efficient task scheduling and execution monitoring in large-scale robot collaboration.

[0024] In summary, the cluster of execution units is composed of multiple robots, with high task execution capability and flexible dynamic adjustment capability. The role of each robot in the cluster is clearly defined according to the functional requirements and task requirements, ensuring accurate control and efficient execution in a dynamically changing working environment.

[0025] Further, before task interpretation and role grouping planning are performed by the task engine, the task engine is constructed, and step S1 of the application includes:

[0026] The task interpretation node is constructed by condition type classification, single-thread decoupling, and deep feature extraction. The role grouping node is constructed by code matching and master-slave grouping. The engineering database is constructed based on historical data by cascading the task interpretation node and the role grouping node. The engineering samples are integrated and supervised training is performed to convergence to determine the task engine. The task engine is embedded and deployed in the task management center.

[0027] In the construction process of the task interpretation node in the embodiments of the application, the tasks are first preliminarily divided by condition type classification. The condition type classification refers to dividing the tasks into different condition categories according to the working environment and conditions of the tasks. For example, inspection tasks, machining tasks, etc. Different task processing procedures are designed for each condition type.

[0028] Next, the tasks are decoupled from complex task modes into multiple single tasks under a single thread for processing by single-thread decoupling, such as decoupling based on single-machine operability of robots, etc., to ensure that each task can be independently executed according to its specific requirements under different conditions.

[0029] The deep feature extraction extracts task features based on a single thread, including task self features of each single thread and interaction and collaboration features between threads. The most representative features are extracted from a large amount of task data to further optimize the accuracy and efficiency of task interpretation.

[0030] In summary, the task interpretation node is constructed according to the above-mentioned bottom-layer execution logic.

[0031] Secondly, the role grouping node is constructed by code matching and master-slave grouping. The code matching refers to matching the unique code of the robot with the task requirements to ensure that each robot can perform corresponding work according to its own ability and task requirements. The master-slave grouping divides the robots into master robots and slave robots according to the nature and complexity of the tasks.

[0032] In an optional embodiment, the master robot is responsible for coordinating the overall progress of task execution, and the slave robot completes specific task operations according to the master status of the master robot. Preferably, the master robot is used as a reference for task progress to coordinate the slave robots and improve the efficiency of collaborative management.

[0033] In summary, the role grouping node is constructed according to the above-mentioned bottom-layer execution logic.

[0034] Next, the task interpretation node is cascaded with the role grouping node, i.e., the role grouping node is placed behind the task interpretation node, and node interaction is established to ensure smooth connection of each link in the task decision, thereby improving the overall operation efficiency of the system.

[0035] Preferably, in order to further improve the intelligent level of the task engine, an engineering database is constructed based on historical data, and engineering samples are integrated. According to the training requirements of the present application, the above-mentioned two bottom logic is used for data logic sequence integration as sample data. Then, the task interpretation node and the role grouping node after cascading are trained to convergence. Preferably, the preset output precision is used as the convergence condition, and the parameters of the task engine are continuously adjusted until the prediction accuracy reaches the best state, and finally the training is completed and the task engine converges.

[0036] Finally, the task engine is embedded and deployed in the task management center. The embedded deployment ensures that the task engine can process task information from the task management center in real time and efficiently, and closely cooperate with the robot cluster.

[0037] In summary, the deployment of the task engine greatly improves the intelligent level of task allocation, so that the entire multi-robot cooperation system can respond quickly and effectively complete tasks when facing complex tasks and dynamic working conditions.

[0038] Further, the task management center is built-in with a robot cluster library, wherein the robot cluster library adopts an encoding form based on a capability matrix; data interaction between the task engine and the robot cluster library is established.

[0039] In the task management center in the embodiment of the present application, a robot cluster library is built-in, which is used to store and manage information of all robots that can perform task scheduling and execution.

[0040] Preferably, the robot cluster library adopts an encoding form based on a capability matrix. The capability matrix in the present application refers to matrix quantization and encoding of robot information according to the functional characteristics and execution capabilities of each robot, so as to facilitate quick retrieval and matching of suitable robots in the process of task allocation and scheduling.

[0041] In one feasible implementation manner of the present application, each row in the capability matrix represents a robot, and each column represents the execution capability or specific attribute of the robot under different task types or working conditions.

[0042] For example, the load capacity of a robot in a carrying task, the stability in a high-temperature environment, and the accuracy in visual processing can be quantified by the encoding form of the capability matrix. The task management center can quickly determine which robots are suitable for performing a specific task, thereby optimizing task scheduling and allocation.

[0043] Meanwhile, in order to ensure efficient data interaction between the task engine and the robot cluster library, a corresponding data interface is established. The task engine obtains robot information in the robot cluster library through the data interface, and based on the encoding of the robot capability matrix, the task engine can select suitable robots according to the task requirements and allocate specific tasks to them.

[0044] In an optional implementation, in the process of data interaction, the information transmission between the task engine and the robot cluster library is bidirectional. Optionally, the task engine sends task allocation information and execution requirements to the robot cluster library, and the robot cluster library feeds back the robot execution results or real-time state information according to the progress state of the task. The bidirectional data flow ensures real-time monitoring and dynamic adjustment in task execution, and can optimize the scheduling of robots according to the feedback information in execution to cope with possible sudden situations or changes in working conditions.

[0045] Further, the task engine interprets the task and plans the role grouping. Step S1 includes:

[0046] The directed task is received, the decoupling rule of the working condition type mapping is triggered by classifying the working condition type, the single-thread decoupling is executed, the multi-thread working condition is determined, the thread internal feature and thread interaction feature are extracted for the multi-thread working condition, the task item cluster is determined, the execution unit matching is performed based on the capability matrix for the task item cluster, the robot code set is determined, each robot code corresponds to a task item, the leading robot selection and the accompanying robot configuration are performed on the robot code set, the grouping optimization is performed by performing the grouping energy efficiency evaluation, and the execution unit cluster is determined.

[0047] In this step, after receiving the directed task, the task engine classifies the task according to the working condition type of the task. The working condition type classification is to determine the working mode of the task according to the task scene and the working condition requirement, for example, inspection type, mechanical processing type, etc. The working condition type determines the execution mode and processing strategy of the task.

[0048] Subsequently, the task engine triggers decoupling rules of the working condition type mapping. In this application, each task type corresponds to a decoupling rule to direct the processing of different working condition types. Through the decoupling rule, the complex multitask environment is separately disassembled into single-threaded tasks that can be processed, simplifying the task execution process and reducing mutual interference. The integrated task is decoupled into single-machine feasibility tasks based on the function orientation of the robot.

[0049] Next, according to the decoupled single-threaded tasks, i.e., the decoupled directional tasks into multiple single-threaded tasks, which may involve multiple parallel subtasks, and then according to the task requirements and execution conditions, it is determined whether the task needs to be processed in parallel, i.e., whether to synchronize, whether to be the same frequency, etc.

[0050] In this process, the task characteristics of each thread are extracted, which may specifically include the spatiotemporal characteristics of the task, the priority of the task execution, the task operation requirements, etc. In addition, the interaction characteristics between threads are analyzed, and the cooperation and coordination between different threads are focused on. Key task items in the task can be identified to form a task item cluster for subsequent task allocation and execution.

[0051] Subsequently, for the task item cluster, the role grouping node of the task engine matches the execution unit using the capability matrix of each robot. In this application, the capability matrix is a quantitative tool for the capabilities of each robot, recording the working capacity of each robot and the performance in specific tasks. Through the capability matrix, the role grouping node can accurately assign task items to suitable robots.

[0052] Optionally, each robot is encoded with a clear identification in the capability matrix, and the encoding of the task item cluster determines the specific tasks that each robot should perform. In this application, the generation of the robot code set helps to accurately manage task allocation and ensures that each robot can efficiently complete its assigned tasks.

[0053] Next, based on the role grouping node, the robot code set is selected for the lead robot and the accompanying robot.

[0054] Specifically, the lead robot relies on the task main line to be responsible for coordinating and controlling the execution process progress of the entire task, while the accompanying robot assists the lead robot to complete specific operations. In one specific embodiment provided by the present application, when selecting the lead robot, the capabilities of the robot, the task requirements, and the complexity of the task are considered to ensure that the selected lead robot can efficiently coordinate the task execution. The configuration of the accompanying robot is adjusted according to the task requirements of the lead robot to ensure that each robot can maximize its capabilities when working together.

[0055] Finally, the execution unit cluster is optimized through grouping energy efficiency evaluation. The purpose of the grouping energy efficiency evaluation is to adjust the roles of robots in tasks and the way of task allocation by analyzing the energy efficiency performance of each robot when executing tasks, and finally determine the optimal execution unit cluster configuration.

[0056] For example, the energy efficiency evaluation based on knowledge reasoning in task execution can randomly set multiple task nodes based on task progress by referring to the same type of task execution logic as the basis for reasoning, analyze the robot execution state of the task nodes, and perform weighted summation according to the importance of the task nodes as the evaluation result. When the evaluation result is not up to standard, grouping adjustment and re-evaluation are performed to select the optimal grouping for a predetermined number of times as the execution unit cluster.

[0057] In summary, the determination of the execution unit cluster not only improves the efficiency of task execution, but also optimizes the energy efficiency of the entire system, so that tasks can be efficiently and stably executed under different working conditions.

[0058] Further, decoupling rules are set, wherein the decoupling rules include a first decoupling rule under a same frequency synchronous working condition, a second decoupling rule under a same frequency asynchronous working condition, a third decoupling rule under a different frequency synchronous condition, and a fourth decoupling rule under a different frequency asynchronous condition.

[0059] In this step, the decoupling rules are first set to ensure the directional analysis and efficient execution of tasks under different working conditions. The purpose of the decoupling rules proposed in this application is to decouple tasks from complex interactions according to the working condition type of the task and the robot execution condition, so that each subtask can be executed independently, reducing interference between each other, thereby improving the overall task execution efficiency.

[0060] Specifically, the decoupling rules include the following four types:

[0061] First, the first decoupling rule under the same frequency synchronous working condition: under the same frequency synchronous working condition, all robots participating in the task are in a synchronous state in time and frequency. At this time, the task engine arranges multiple robots to execute similar tasks in parallel under the same working frequency according to the preset first decoupling rule. Since the synchronization between robots is strong, the execution efficiency can be improved by sharing resources and information. Specifically, the first decoupling rule mainly functions to coordinate the task order and execution time of each robot.

[0062] Second, the second decoupling rule under the same frequency asynchronous condition: when the robot working frequency is the same, but the timing of executing tasks is different (i.e. asynchronous execution), the task engine triggers the second decoupling rule. The purpose is to adjust the start time and execution order of the task, so that the robot can execute different tasks at the same frequency without conflict. The second decoupling rule mainly resolves the frequency interaction through time in this case, ensuring that each robot can efficiently complete the task according to its own rhythm.

[0063] Third, the third decoupling rule under the same frequency synchronous condition: in the same frequency synchronous condition, the robots are in different working frequencies, but the task execution is synchronous. This means that each robot works independently and cooperatively according to its own frequency, but the execution timing is the same. The role of the third decoupling rule is to coordinate the frequency differences between robots, so that the resource conflicts caused by different frequencies can be minimized when executing tasks synchronously. The third decoupling rule will adapt to the frequency and task content of each robot to ensure that the task can be completed cooperatively.

[0064] Fourth, the fourth decoupling rule under the same frequency asynchronous condition: in the same frequency asynchronous condition, the robots not only have different frequencies, but also have different execution order and timing of tasks. It usually brings greater coordination difficulty. Therefore, the fourth decoupling rule needs more precise control. This rule controls the execution of tasks with different frequencies and timing through scheduling algorithms, so that each robot can independently and in parallel execute tasks while avoiding conflicts in frequency and timing. The decoupling rule dynamically adjusts the allocation and execution time of tasks to ensure that robots can efficiently complete their tasks independently and cooperatively.

[0065] In summary, it can flexibly cope with different types of conditions and process them according to the characteristics of the task to ensure efficient and stable execution of the task, while optimizing the overall performance of the robot cluster. Each decoupling rule can effectively reduce interference and reasonably schedule tasks according to the type of working condition, thereby improving the coordination and task execution efficiency of multiple robots.

[0066] S2: Establish a task communication system based on the execution unit cluster through a 5G private network, and deploy dynamic management components at the task edge.

[0067] In this step, first, a task communication system based on the execution unit cluster is established through a 5G private network. As a high-speed, low-latency private communication network, 5G private network can provide stable and efficient data transmission channels between multiple robot clusters.

[0068] The task communication system is proposed, which is the basis for realizing the cooperative work of multiple robots. Specifically, the communication network deployed is interactively associated with the local communication topology, that is, the communication interaction of the execution unit cluster is established. Through the communication system, each robot in the execution unit cluster can exchange task data, state information and control instructions in real time. The self-control management of the robot cluster based on the task is ensured, and when there are multiple tasks, each robot cluster can realize cluster self-control without interference.

[0069] At the same time, the use of 5G private network ensures high reliability and low delay of communication, especially in complex and dynamic working environment, which can ensure rapid transmission of task instructions between robots, greatly improving the response speed and cooperation ability of task execution.

[0070] Specifically, in the process of establishing the task communication system, the robot cluster is interconnected through the 5G private network when executing the task, and the data stream and control instruction can be transmitted in real time. Guided by the existence of the task, a temporary private network interconnection interaction is established for the self-control of the robot cluster based on the task, and the task communication system is released after the task is completed.

[0071] Next, the data of the task communication system is converged to the task edge and the dynamic management component is deployed. The task edge is an intermediate architecture between the task management center and the robot, which is the edge side of the robot, and it converges the task execution information, state data, sensor data, etc. of the robot cluster.

[0072] In the technical route of the present application, the task management center plans and allocates based on the task, and the task edge takes over and coordinates the supervision of the robot cluster in execution.

[0073] Preferably, in the task edge, the dynamic management component is deployed for real-time management and scheduling. The role of the dynamic management component is to dynamically adjust and optimize the task execution of the robot cluster. That is, through the task edge, the planned robot cluster planning is taken over, bidirectional data interaction is carried out according to the communication interaction between the task edge and the robot central control, and the dynamic management component is assisted. When an abnormality or deviation occurs in the task execution process, the dynamic management component can timely adjust the task allocation strategy, optimize the resources and path of task execution, and ensure that the task can be completed smoothly according to the predetermined target. Accurately identify the bottlenecks and problems in robot execution, respond quickly and adjust the execution strategy to ensure the efficiency and stability of the entire task flow.

[0074] In summary, the deployment of the task communication system and the dynamic management component based on the 5G private network enables the multiple robot cluster to respond to environmental changes and task requirements in real time when executing complex tasks, achieving more intelligent and efficient task execution.

[0075] Further, the dynamic management component is deployed, and the step S2 of the present application comprises:

[0076] A task coordination port is deployed, wherein the task coordination port performs edge driving and data stream reception; for the task item cluster, key autonomous task points and interactive task points are mined as risk control management task points; for the risk control management task points, a multi-working-condition-regulation parallel branch is constructed, wherein the parallel branch is extensible and at least includes a first branch and a second branch; the dynamic management component is built with the task coordination port and the parallel branch, and the dynamic management component is deployed at the task edge.

[0077] In this step, a task coordination port is first deployed, which is responsible for performing edge driving and data stream reception, and performing bidirectional interaction with the robot cluster.

[0078] Specifically, the role of edge driving is to send task execution instructions to the robot cluster based on the execution unit cluster planned by the task management center, to control the task execution process of the robot. Data stream reception refers to the task coordination port receiving state feedback, sensor data, real-time task execution progress and other information from the robot cluster for real-time monitoring and data analysis. The deployment of the task coordination port ensures the coordination and real-time performance of task execution, and can respond to changes in the state of the robot cluster and adjustments in task requirements in a timely manner.

[0079] Next, the task coordination port mines key task points according to the task item cluster. The task item cluster is a plurality of task units determined by the task engine, which includes various specific operations required for task execution. Some of the task points in the task unit are crucial to the smooth execution of the task, so they need to be focused on. Through analysis of the task item cluster, key autonomous task points and interactive task points are mined.

[0080] Among them, the autonomous task point refers to an independent execution task that does not depend on other task points, i.e., a single robot execution task key point; while the interactive task point is a task point that has a dependent relationship between multiple task points and needs to be completed collaboratively. Taking it as a risk control management task point is a key step to ensure that risk assessment and control can be performed in real time during task execution.

[0081] That is, through screening, task exception determination is performed for task key points, and on the basis of ensuring task main line standardization, fault-tolerant management is performed on the micro-impact task part.

[0082] Further, for the risk control management task point, a parallel branch of multi-working condition regulation is constructed. The construction of the parallel branch is to flexibly adjust and optimize the task execution in multiple working conditions. The parallel branch plays a role in adjusting and correcting during task execution, ensuring that the task can proceed as expected under different working conditions.

[0083] Among them, the parallel branch has scalability, which means that the branch can be dynamically increased according to task requirements and execution conditions, at least including the first branch and the second branch. The first branch and the second branch correspond to different task execution strategies, the first branch focuses on robot off-axis control working condition, and the second branch focuses on robot fault working condition, corresponding to different abnormal conditions, so as to improve the pertinence of decision regulation.

[0084] Finally, the task coordination port is taken as the front, and the first branch and the second branch are connected, and the architecture of the dynamic management component is built.

[0085] Further preferably, by calling historical regulation records, data is arranged based on the above architecture, that is, including original data samples-exception judgment samples-branch output samples, based on which supervised training under sample driving is performed on the architecture until the convergence condition is met, as the completed dynamic management component.

[0086] In the specific implementation process, during task execution, the dynamic management component can analyze the task progress in real time, adjust and optimize according to different working conditions. When detecting execution exceptions or task deviation from the predetermined target, the dynamic management component will adjust the task execution strategy according to the feedback information of the risk control management task point, to ensure that the task can be completed smoothly.

[0087] By deploying the dynamic management component at the task edge, the timely response and efficient scheduling of task execution are ensured, and at the same time, the flexibility and reliability of the task in complex environment are also ensured.

[0088] Further, the parallel branch of multi-working condition regulation is constructed, and the step S2 of the present application comprises:

[0089] For the robot off-axis control working condition, the first branch is built by robot self-driving adjustment and collaborative adjustment; for the robot fault working condition, the second branch is deployed based on the bottom logic of fault blocking-local reorganization-task dynamic redistribution-safety isolation zone activation.

[0090] In this step, first, for the robot off-axis control working condition, the first branch is built by robot self-driving adjustment and collaborative adjustment.

[0091] Among them, the robot off-axis control working condition refers to the situation that the robot deviates from the predetermined trajectory or task target during task execution, which may be caused by factors such as external environment changes, robot state changes or task requirement changes.

[0092] To deal with this working condition, the core of the first branch is to adjust the robot through self-driving adjustment, so that the robot can automatically adjust its running state, trajectory or posture to restore normal working mode. Specifically, the robot self-driving adjustment monitors its position, posture and motion state in real time through the built-in sensors and control system of the robot, and automatically starts the adjustment mechanism to correct the robot motion path or direction when off-axis deviation is detected.

[0093] At the same time, cooperative adjustment refers to the adjustment or coordinated action of other robots according to the state of the master robot in multi-robot cooperative tasks. For example, if a robot deviates from the axis, other robots may adjust their positions or share part of the task to ensure the smooth progress of the overall task. Through self-driving adjustment and cooperative adjustment, the robot off-axis working condition can be effectively dealt with to ensure the continuous execution of the task.

[0094] Next, for the robot failure working condition, the second branch is deployed, and its underlying logic includes fault blocking, local grouping reconstruction, task dynamic redistribution and safety isolation zone activation. Robot failure working condition usually refers to the situation that the robot fails to continue executing the task due to device failure, sensor failure or other system problems during task execution.

[0095] To deal with this working condition, the second branch first performs fault blocking, that is, when a robot failure is detected, the faulty robot is immediately stopped from participating in task execution to prevent it from continuing to affect the progress of the entire task.

[0096] Subsequently, the local grouping reconstruction step starts, which redistributes tasks and adjusts the robot grouping structure. If a robot that was originally responsible for a task cannot continue to execute due to failure, its task is redistributed to other robots according to task requirements. Task dynamic redistribution further ensures the flexibility and continuity of the task, dynamically adjusting the task executor according to task progress and robot capability to ensure that the task can be restored to normal in the shortest time.

[0097] Finally, a safety isolation zone is activated around the faulty robot, which aims to ensure that the problem of the faulty robot does not affect the safety of other robots and the entire working environment. The safety isolation zone restricts the interference of the faulty robot to the surrounding robots through physical or virtual means, ensuring that other robots can continue to execute tasks in a safe environment.

[0098] In summary, the bottom logic principles of the first branch and the second branch are set, and the branch deployment and further supervision training are carried out based on this.

[0099] Through the above deployment principle, when facing different robot failures or deviation working conditions, it can quickly respond and make effective adjustments. The first branch ensures that the robot can self-adjust when it deviates from the axis and cooperate with other robots, and the second branch provides an efficient fault isolation and task recovery mechanism when the robot fails, ensuring that the task execution does not interrupt and is completed efficiently.

[0100] S3: Based on the task communication system and the dynamic management component, take over the execution unit cluster to drive control and track supervision with robot coding, wherein the reassignment and safety isolation control of reorganization under abnormal working condition execution robot task control adjustment and fault blocking.

[0101] In this step, the task edge end takes over the execution unit cluster and drives the control of task execution based on the task communication system and the dynamic management component.

[0102] Among them, the execution unit cluster is composed of multiple robots, which is responsible for performing various tasks in the task. The task communication system, as the basis for multi-robot collaboration, ensures that robots can transmit task instructions, state information and other key data in real time. The dynamic management component is responsible for dynamically adjusting and optimizing task execution, ensuring that it can flexibly respond to various environmental changes and task demand changes during task execution.

[0103] Specifically, robot coding is used as the unique identifier of each robot, and the task of each robot is accurately managed through robot coding. Robot coding is not only used for task allocation, but also can track the state and performance of each robot during task execution, ensuring that the role and responsibility of each robot in the cluster is clear.

[0104] On this basis, the dynamic management component embedded in the task edge end accurately controls the robot through drive control, and realizes the autonomous communication interaction of the task cluster according to the task communication system. Drive control refers to sending instructions to control the movement, working mode and task execution of each robot according to the progress of the task and the real-time state of the robot. For example, when the task requires the robot to perform a certain operation, according to the current position, working state and other information of the robot, the action is accurately controlled, so that the robot can work efficiently on the predetermined track.

[0105] Further, through tracking supervision, the task management system can obtain real-time data of the robot when executing the task at any time, including position, speed, task progress and other information. If it is detected that a robot deviates from the task plan or appears abnormal during execution, it will respond immediately.

[0106] Preferably, during the task execution process, special attention is paid to the occurrence of risk control management task points. When abnormal conditions are detected, such as the robot deviating from the predetermined trajectory, robot failure or external interference affecting task execution, task control adjustment is initiated, that is, based on the first branch in the dynamic management component, the execution mode of the task is adjusted in real time according to the current task demand and the state of the robot, to ensure that the task can be performed as expected.

[0107] If a robot failure or other major abnormality occurs during task execution, a fault blocking mechanism is triggered according to the second branch. The fault blocking mechanism will immediately stop the task execution of the faulty robot and reassign tasks according to the task demand. At this time, reassignment is further performed by reconfiguring the group to reassign the tasks originally responsible by the faulty robot to other healthy robots, ensuring the continuous performance of the task.

[0108] Optionally, for the current robot cluster, the cluster apportionment of the task of the faulty robot is performed, or a new robot is called to take over the task and the new robot is integrated into the task communication system.

[0109] Finally, safety isolation control is initiated, that is, when the faulty robot may pose a threat to other robots or tasks, the faulty robot is isolated from other robots by setting up a safety isolation zone to prevent the impact of the fault on the entire task execution.

[0110] In summary, it is ensured that when any abnormal condition occurs, the robot cluster can flexibly respond and ensure efficient completion of the task.

[0111] Further, the robot code is driven and controlled and tracked and supervised. The step S3 of the present application comprises:

[0112] The execution unit cluster is deployed to the task coordination port of the dynamic management component; based on the task coordination port and taking the robot code as the communication identifier, the driving control of the task item cluster is performed; the job flow data is determined according to the robot central control system, wherein the job flow data is identified by a space-time code; the job flow data is returned, the data node of the risk control management task point is located, and the abnormal condition evaluation under the ternary data sequence is performed in combination with the front and rear data nodes to determine the evaluation result.

[0113] In this step, the execution unit cluster is first deployed to the task coordination port of the dynamic management component. The task coordination port is responsible for receiving and processing task information from the execution unit cluster, controlling and driving the robot based on the execution unit cluster, and at the same time, receiving the task data flow of the robot central control.

[0114] In the present application, the execution unit cluster is composed of multiple robots, which exchange real-time data and transmit instructions with the task coordination port through the task communication system during task execution. After the execution unit cluster is decentralized to the task coordination port, the task coordination port can fully control the state and task progress of all robots in the cluster, ensuring the coordination and flexibility of the entire system.

[0115] Next, the task coordination port starts to perform driving control of the task item cluster based on the task communication system. The task communication system is the basis for the coordinated work of multiple robots, which ensures that task instructions can be quickly and accurately transmitted from the task management system to each robot, and realizes autonomous communication of the cluster, avoiding communication interference from other ports.

[0116] Specifically, based on the task coordination port, the task item cluster is accurately scheduled and driven according to the code of the robot, i.e. the unique identifier of each robot as a communication identifier.

[0117] Further, during task execution, the work flow data is determined through the robot's central control system. The robot's central control system records its running data and performs space-time code identification as the robot's work flow data, i.e. the dynamic data flow of the robot during task execution, including task progress, state, position, etc.

[0118] Through space-time code identification, i.e. the encoding of time and space information during task execution, the position, state and specific time of task execution of the robot can be identified. Through space-time code, the position of the robot and the task progress can be tracked in real time, and accurate task scheduling and management can be performed.

[0119] Subsequently, based on the task communication system, the work flow data is transmitted back to the task edge to be received by the task coordination port in the dynamic management component. Next, by positioning the data node of the risk management task point, the key node in task execution is determined. The risk management task point is a point that needs special attention during task execution, which may need to be assessed and adjusted due to task complexity, environmental changes or changes in robot state, i.e. matching the work flow data position based on the risk management node as the data node.

[0120] Further, the preposed data node and the postposed data node are combined to perform abnormality evaluation of the working condition based on the ternary data sequence. The ternary data sequence consists of the data part of the data node, the data part of the previous time sequence node of the data node, and the data part of the next time sequence node of the data node. Based on the single node data state and the time sequence data trend, the decision accuracy can be improved.

[0121] Further, the working condition anomaly is identified for the ternary data sequence. The working condition anomaly evaluation timely discovers potential problems or abnormal conditions in task execution by analyzing the relationship between each data node. A specific determination method provided in the application is: determining according to the difference between the expected data of the execution unit cluster and the ternary data sequence, when it does not meet the expectation or the difference is greater than the task allowed deviation, it is determined as a working condition anomaly.

[0122] For example, if the robot deviates from the task trajectory, the task progress lags behind, or the equipment fails, etc., the anomaly is quickly identified, and the response is made according to the evaluation result, and the corresponding control measures are taken.

[0123] Specifically, if it is a robot off-axis control working condition, the first branch is triggered to make a decision, and if it is a robot failure working condition, the second branch is triggered to make a decision. For example, adjusting the robot task allocation, re-grouping or starting the fault handling mechanism.

[0124] In summary, it can ensure that in a complex multi-robot cooperation environment, the task can be smoothly executed as expected, and timely adjustment and processing are made when encountering abnormal working conditions, ensuring efficient and safe completion of the task.

[0125] Further, if the evaluation result is a robot off-axis control working condition, the first branch is triggered to determine the first control strategy; if the evaluation result is a robot failure working condition, the second branch is triggered to determine the second control strategy, wherein the task dynamic re-allocation mode is based on replacement allocation of a new robot or re-allocation of the original group.

[0126] In this step, different control strategies are taken to deal with different task execution problems according to the results of working condition anomaly evaluation.

[0127] First, when the evaluation result is a robot off-axis control working condition, the first branch is triggered and the first control strategy is determined. The off-axis control working condition usually refers to the situation that the robot deviates from the predetermined path or target during task execution, and the specific cause may be caused by task environment changes, robot operation errors, etc.

[0128] Specifically, the core goal of the first control strategy is to restore the accuracy and stability of the robot execution through self-driving adjustment and cooperative adjustment of the robot. The first branch takes the deviation degree as the adjustment target to ensure that it returns to the normal working state. At the same time, the cooperative robot of this task node can also be adjusted according to the task demand, and the deviation is corrected through cooperative adjustment, so as to avoid the influence of the off-axis of a single robot on the entire task execution process.

[0129] Secondly, if the evaluation result is a robot failure condition, the second branch is triggered and the second control strategy is determined. Robot failure condition usually refers to the situation that the robot cannot continue to complete the task due to equipment damage, sensor failure, control system failure, etc. during task execution. It may seriously affect the task progress, and even cause the task to be interrupted.

[0130] To solve this problem, the implementation steps of the second control strategy include fault blocking, local group reconstruction, task dynamic redistribution and safety isolation control.

[0131] Firstly, the fault robot is blocked to ensure that it stops executing the task and does not further affect other parts of the system.

[0132] Subsequently, local group reconstruction is performed to redistribute the task from the faulty robot to other robots. Task dynamic redistribution refers to the real-time redistribution of tasks to other robots according to the situation that the faulty robot cannot continue to execute the task. There are two specific redistribution methods: one is replacement allocation based on new robots, that is, a standby robot is introduced to replace the faulty robot to continue to complete the task; the other way is to redistribute the original group, that is, the task of the faulty robot is allocated to other robots in the same group, and other robots take over the task originally completed by the faulty robot.

[0133] In summary, the robot failure problem can be effectively addressed to ensure that the task can continue and reduce the impact of failure on task execution. The above decision-making process is automatically driven by the training mechanism of the dynamic management component.

[0134] In addition, safety isolation control will be started when necessary to ensure that the faulty robot no longer poses a safety threat to other robots or tasks, and to ensure the safety and stability of task execution.

[0135] The multi-robot synchronous control method provided by the present application has the following technical effects:

[0136] 1. Task-driven intelligent grouping: Construct a task engine, analyze the task through condition classification, single-thread decoupling and feature extraction, group master and slave roles according to the robot capability matrix, and optimize the execution unit cluster through energy efficiency evaluation. Realize accurate matching of task and robot capability, avoid resource waste, and improve multi-robot cooperation efficiency and task execution success rate.

[0137] 2. 5G private network communication system construction: Based on 5G private network, construct task communication system, deploy dynamic management component, realize real-time data aggregation and edge driving control of execution unit cluster. Ensure low delay and high reliability of multi-robot communication, support fast task instruction issuing and state feedback, and enhance system response real-time performance.

[0138] 3. Dynamic management and risk response: The dynamic management component sets the task coordination port and the multi-working condition control parallel branch. For robot off-axis, failure and other abnormalities, it respectively executes self-driven cooperative adjustment or fault blocking, reorganization and other strategies. Through modular risk management, it quickly responds to abnormal working conditions, reduces the impact of task interruption, and ensures system stability and safety. Coding drive and closed-loop supervision: The robot code is used as the identifier to drive task execution, and the task flow data is returned for space-time code tracking. Based on the risk management task point, the three-dimensional data sequence abnormality is evaluated and the strategy is triggered. It realizes full-process task monitoring and precise traceability, ensures the controllability of multi-robot operation, and improves the overall reliability of the system.

[0139] Embodiment two: based on the same inventive concept as the multi-robot synchronization control method in the preceding embodiment, as shown in Figure 2 The application provides a multi-robot synchronization control system, which comprises:

[0140] A task planning unit 11 is used to upload a directional task to a task management center, and a task engine is used for task interpretation and role grouping planning to determine an execution unit cluster. The role grouping is performed in a function-oriented master-slave cooperation mode, and the robot code is used as the execution unit.

[0141] A communication deployment unit 12 is used to establish a task communication system based on the execution unit cluster through a 5G private network, and a dynamic management component is deployed at the task edge.

[0142] A control supervision unit 13 is used to take over the execution unit cluster based on the task communication system and the dynamic management component, drive control and tracking supervision are performed based on the robot code, and abnormal working condition execution robot task participation control adjustment, reassignment and safety isolation control under fault blocking and reorganization are performed.

[0143] Further, the task planning unit 11 is used to perform the following steps: working condition type classification-single thread decoupling-depth feature extraction, construction of a task interpretation node; code matching-master-slave grouping, construction of a role grouping node; cascading the task interpretation node and the role grouping node, constructing an engineering database based on historical data, integrating engineering samples and supervising training to convergence, determining the task engine, and embedding the task engine in the task management center.

[0144] Further, the task management center is built-in with a robot cluster library, wherein the robot cluster library adopts a coding form based on a capability matrix; data interaction between the task engine and the robot cluster library is established.

[0145] Further, the task planning unit 11 performs the following steps: receiving the directional task, triggering the decoupling rule of the working condition type mapping by performing working condition type classification, executing single-thread decoupling, and determining multi-thread working conditions; for the multi-thread working conditions, performing thread internal feature and thread interaction feature extraction, and determining a task item cluster; for the task item cluster, performing execution unit matching based on the capability matrix, and determining a robot code set, wherein each robot code corresponds to a task item; for the robot code set, performing leading robot selection and accompanying robot configuration, performing grouping optimization by performing grouping energy efficiency evaluation, and determining the execution unit cluster.

[0146] Further, a decoupling rule is set, wherein the decoupling rule includes a first decoupling rule in a same frequency synchronous working condition, a second decoupling rule in a same frequency asynchronous working condition, a third decoupling rule in a different frequency synchronous working condition, and a fourth decoupling rule in a different frequency asynchronous working condition.

[0147] Further, the communication deployment unit 12 is used to perform the following steps: deploying a task overall port, wherein the task overall port performs edge driving and data flow receiving; for the task item cluster, mining key autonomous task points and interaction task points as risk control management task points; for the risk control management task points, constructing a parallel branch of multi-working condition regulation, wherein the parallel branch is extensible and at least includes a first branch and a second branch; with the task overall port and the parallel branch, building the dynamic management component, and deploying the dynamic management component at the task edge.

[0148] Further, the communication deployment unit 12 performs the following steps: for a robot off-axis control working condition, building a first branch with robot self-driving adjustment and cooperative adjustment; for a robot fault working condition, deploying a second branch with fault blocking-local grouping reconstruction-task dynamic redistribution-safety isolation zone activation as the underlying logic.

[0149] Further, the control supervision unit 13 performs the following steps: downgrading the execution unit cluster to the task overall port of the dynamic management component; taking the task communication system as a benchmark, based on the task overall port, taking the robot code as a communication identifier, and performing driving control of the task item cluster; according to the robot central control system, determining job flow data, wherein the job flow data is identified with a space-time code; returning the job flow data, positioning the data node of the risk control management task point, and jointly performing working condition abnormality evaluation under the ternary data sequence with the front data node and the rear data node, and determining the evaluation result.

[0150] Further, if the evaluation result is a robot off-axis control working condition, a first branch is triggered to determine a first regulation strategy; if the evaluation result is a robot failure working condition, a second branch is triggered to determine a second regulation strategy, wherein the task dynamic re-distribution mode is replacement distribution based on a newly added robot or original marshalling re-distribution.

[0151] The person skilled in the art can clearly understand the multi-robot synchronous control method and system in the embodiment according to the foregoing detailed description of the multi-robot synchronous control method. As the device disclosed in the embodiment corresponds to the method disclosed in the embodiment, the description is relatively simple, and the related parts can be referred to the method part.

[0152] The above description of disclosed embodiments enables a person skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A multi-robot synchronous control method, characterized in that: The method comprises: Upload the directional tasks to the task management center, use the task engine to interpret the tasks and plan the role grouping, and determine the execution unit cluster. Among them, the role grouping is carried out based on the function-oriented master-slave collaboration, and the robot code is used as the execution unit; Establish a task communication system based on the execution unit cluster using a 5G private network, converge at the task edge and deploy dynamic management components; Based on the task communication system and the dynamic management component, the execution unit cluster is taken over and drive control and tracking supervision are performed using robot coding. Among them, the robot task parameter control adjustment and grouping reconstruction and redistribution and safety isolation management under abnormal working conditions and fault blocking are performed.

2. A multi-robot synchronous control method according to claim 1, characterized in that: Before using the task engine to interpret tasks and plan role groups, the construction of the task engine includes: Build task interpretation nodes based on working condition classification, single-thread decoupling, and deep feature extraction; Use coding matching - master-slave grouping to build role grouping nodes; The task interpretation node and the role grouping node are cascaded, an engineering database is constructed based on historical data, engineering samples are integrated and training is supervised until convergence, the task engine is determined, and the task engine is embedded and deployed in the task management center.

3. A multi-robot synchronous control method according to claim 2, characterized in that: The task management center has a built-in robot cluster library, wherein the robot cluster library adopts a coding form based on the capability matrix; Establish data interaction between the task engine and the robot cluster library.

4. A multi-robot synchronous control method according to claim 3, characterized in that: Use the mission engine to interpret missions and plan role formations, including: receiving the directed task, classifying the working condition type, triggering the decoupling rule of the working condition type mapping, executing single-thread decoupling, and determining the multi-thread working condition; For the multi-threaded working condition, extract thread internal features and thread interaction features to determine task item clusters; For the task item cluster, perform execution unit matching based on the capability matrix to determine a robot code set, wherein each robot code corresponds to a task item; For the robot code set, the leading robot selection and the accompanying robot configuration are performed, the grouping optimization is performed by performing grouping energy efficiency evaluation, and the execution unit cluster is determined.

5. A multi-robot synchronous control method according to claim 4, characterized in that: Decoupling rules are set, wherein the decoupling rules include a first decoupling rule under the same-frequency synchronous working condition, a second decoupling rule under the same-frequency asynchronous working condition, a third decoupling rule under the different-frequency synchronous working condition, and a fourth decoupling rule under the different-frequency asynchronous working condition.

6. A multi-robot synchronous control method according to claim 4, characterized in that: Deploy dynamic management components, including: Deploy a task coordination port, wherein the task coordination port performs edge driving and data stream reception; For the task item cluster, key autonomous task points and interactive task points are mined as risk control management task points; For the risk control management task point, a parallel branch for multi-operating condition control is constructed, wherein the parallel branch is expandable and includes at least a first branch and a second branch; The dynamic management component is constructed using the task coordination port and the parallel branch, and the dynamic management component is deployed at the task edge.

7. A multi-robot synchronous control method according to claim 6, characterized in that: Construct parallel branches for multi-condition control, including: For the off-axis control condition of the robot, the first branch is built by using the robot's self-drive adjustment and coordinated adjustment; In response to robot failure conditions, the second branch is deployed with the underlying logic of fault blocking - local grouping reconstruction - dynamic task redistribution - safety isolation zone activation.

8. A multi-robot synchronous control method according to claim 1, characterized in that: Use robot coding for drive control and tracking supervision, including: Delegating the execution unit cluster to the task coordination port of the dynamic management component; Based on the task communication system, the task coordination port and the robot code as the communication identifier, the driving control of the task item cluster is executed; Determining operation flow data according to the central control system of the robot, wherein the operation flow data is identified by a time-space code; The operation flow data is transmitted back, the data node of the risk control management task point is located, and the preceding data node and the following data node are combined to perform an abnormal working condition assessment under the ternary data sequence to determine the assessment result.

9. A multi-robot synchronous control method according to claim 8, characterized in that: If the evaluation result is an off-axis control condition of the robot, trigger the first branch and determine the first control strategy; If the evaluation result is a robot failure condition, the second branch is triggered and a second control strategy is determined, wherein the task dynamic redistribution method is based on replacement allocation of newly added robots or redistribution of the original grouping.

10. A multi-robot synchronous control system, characterized in that: For executing a multi-robot synchronous control method according to any one of claims 1 to 9, the system comprises: The task planning unit is used to upload the targeted tasks to the task management center, interpret the tasks and plan the role grouping using the task engine, and determine the execution unit cluster. The role grouping is performed based on function-oriented master-slave collaboration, and the robot code is used as the execution unit. A communication deployment unit, configured to establish a task communication system based on the execution unit cluster using a 5G private network, converge at the task edge, and deploy dynamic management components; The control and supervision unit is used to take over the execution unit cluster based on the task communication system and the dynamic management component, and perform drive control and tracking supervision with robot coding, wherein the robot task parameter adjustment and grouping reconstruction and redistribution and safety isolation management under abnormal working conditions and fault blocking are performed.

Images