A control method and system for multi-machine collaborative operation based on inductive computing and control technology

Abstract

This application relates to the field of information technology and discloses a control method and system for multi-robot collaborative operation based on sensing and computational control technology. The method includes: acquiring the task granularity distribution and resource requirement matrix, classifying them using a classification algorithm to obtain an initial task segmentation feature set; generating a preliminary task allocation scheme based on the optimal segmentation point set and considering the heterogeneity of node computation, and dynamically updating the task segmentation feature set through real-time sensing technology; generating a real-time task state set through a group information sharing mechanism, optimizing the task allocation scheme using a dynamic programming algorithm to ensure that tasks are allocated reasonably according to priority and node resources; and updating the collaborative state matrix through a real-time communication protocol to ensure the accuracy and timeliness of the multi-node collaborative execution results. This invention can dynamically adjust task allocation in complex environments, optimize resource utilization, and improve the collaborative efficiency of multi-robot systems, and can be widely applied in fields such as industrial production, warehousing and logistics, and disaster relief.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a multi-robot collaborative operation control method and system based on sensing and control technology, which is especially applicable to multi-robot task allocation and collaborative operation optimization in fields such as robot systems, automated production, logistics warehousing, and disaster relief. Background Technology

[0002] With the continuous development of automation technology and robotic systems, especially in the field of multi-robot collaborative operations, how to efficiently allocate and collaborate tasks has become a key technical issue. Multi-robot systems (MRS) are widely used in complex environments such as industrial production, warehousing and logistics, and disaster relief. The tasks in these fields are typically highly complex and dynamic, and the collaborative execution between robots needs to be adjusted in real time under different environmental conditions and task requirements.

[0003] Currently, most existing multi-robot task allocation methods rely on static presets or rule-based scheduling strategies, lacking the ability to dynamically adjust to factors such as task complexity, resource requirements, and robot heterogeneity. Traditional methods often struggle to achieve efficient task allocation when faced with complex factors such as frequent task changes, differences in robot computing power and resources, and variations in communication bandwidth, leading to low task execution efficiency or resource waste.

[0004] To address these issues, researchers have proposed various task allocation methods based on intelligent algorithms in recent years, such as genetic algorithms and particle swarm optimization algorithms. However, these methods typically require long computation times and are difficult to respond to dynamically changing tasks and environments in real time, which limits their widespread use in practical applications.

[0005] Furthermore, existing technologies also have problems with information sharing and coordination mechanisms. Although some methods attempt to optimize task scheduling through information sharing, in real-world environments, communication between robots is often affected by factors such as bandwidth and latency, leading to untimely information synchronization and consequently impacting task execution efficiency and collaborative effectiveness.

[0006] Therefore, how to perform real-time and dynamic task allocation in multi-robot systems, while considering multiple factors such as task complexity, robot heterogeneity, resource requirements, and real-time communication, is a key challenge in current technological development. The main objective of this invention is to provide a control method and system for multi-robot collaborative operations based on sensing and computational control technology. By real-time sensing and dynamic optimization of task allocation schemes, it effectively solves the problems of low task allocation efficiency, severe resource waste, and lagging information sharing in existing technologies for multi-robot systems. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention adjusts the task allocation scheme in real time based on dynamic factors such as changes in tasks, the computing power of robot nodes, and changes in communication bandwidth, thereby improving the collaborative efficiency of multi-robot systems.

[0008] In a first aspect, this application provides a control method for multi-machine collaborative operation based on sensing and control technology, the method comprising:

[0009] Step S1: Obtain the task granularity distribution and resource requirement matrix, and use a classification algorithm to classify the task granularity distribution and resource requirement matrix to obtain the initial task segmentation feature set;

[0010] Step S2: Extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and use cluster analysis to determine the optimal segmentation point set;

[0011] Step S3: Based on the optimal split point set, and combined with the heterogeneity of node computation, generate a preliminary task allocation scheme;

[0012] Step S4: Update the task granularity distribution and resource requirement matrix through real-time perception technology to obtain a dynamically adjusted task segmentation feature set;

[0013] Step S5: Determine the real-time task state set through the group information sharing mechanism, and use the dynamic programming algorithm to optimize the subtask dependency weights and bandwidth allocation weights to generate an optimized task allocation scheme;

[0014] Step S6: Extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through the real-time communication protocol to obtain the collaborative execution results.

[0015] In conjunction with the first aspect, in the first implementation of the first aspect of this application, the process of updating the cooperative state matrix through the real-time communication protocol further includes:

[0016] Add a local timestamp to the collaborative status data reported by each node;

[0017] For each item in the cooperative state matrix, calculate the state lag value based on the difference between the most recent timestamp corresponding to each item and the current system clock.

[0018] If the state lag value exceeds the set lag tolerance threshold, a fast recovery mechanism is triggered, which requests the target node to rebroadcast its current state information to cover outdated data.

[0019] During the matrix update process, a credibility weighted calculation strategy is adopted for the state information from multiple nodes, where the node credibility is calculated based on its past state reporting frequency, communication success rate and latency jitter statistics.

[0020] By employing the aforementioned timestamp control and credibility-weighted update strategies, data inconsistencies in the collaborative state matrix caused by asynchronous or delayed processes are reduced, thereby improving the real-time performance of task scheduling decisions.

[0021] Secondly, this application provides a multi-machine collaborative operation control system based on sensing and control technology, the system comprising:

[0022] The initial task segmentation unit is used to obtain the task granularity distribution and resource requirement matrix. A classification algorithm is used to classify the task granularity distribution and resource requirement matrix to obtain the initial task segmentation feature set.

[0023] The optimal segmentation unit is determined to extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and cluster analysis is used to determine the optimal segmentation point set.

[0024] The preliminary task allocation unit is used to generate a preliminary task allocation scheme based on the optimal set of split points and the heterogeneity of node computation.

[0025] The dynamic task segmentation unit is used to update the task granularity distribution and resource requirement matrix through real-time sensing technology to obtain a dynamically adjusted task segmentation feature set.

[0026] The task allocation unit is optimized to determine the real-time task status set through a group information sharing mechanism, and to optimize the subtask dependency weights and bandwidth allocation weights using a dynamic programming algorithm to generate an optimized task allocation scheme.

[0027] The collaborative execution unit is used to extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through a real-time communication protocol to obtain the collaborative execution results.

[0028] Compared with the prior art, the beneficial effects of the present invention are at least as follows:

[0029] 1. The technical solution provided in this application can achieve accurate task allocation based on factors such as task complexity, resource requirements, and node computing capabilities through classification algorithms, cluster analysis, and dynamic adjustment mechanisms. Especially when facing multi-robot collaborative operations, it can quickly respond to task changes, optimize task allocation, and improve the overall efficiency of task execution.

[0030] 2. The technical solution provided in this application introduces real-time perception technology and dynamic programming optimization algorithm, enabling the present invention to adjust the task allocation scheme in a timely manner according to changes in the environment and task progress; it can flexibly adapt to different working scenarios, handle different task priorities, node computing power and bandwidth limitations, and significantly improve the adaptability of the system.

[0031] 3. In the technical solution provided in this application, through dynamic perception and optimization algorithms, the task granularity distribution can be flexibly adjusted according to the current resource and bandwidth conditions, thereby making more reasonable use of the system's computing resources and communication bandwidth; this refined resource scheduling can reduce resource waste and improve the overall resource utilization efficiency of the system.

[0032] 4. In the technical solution provided in this application, by updating the real-time communication protocol and the collaborative state matrix, each node can maintain synchronization during collaborative operation and provide timely feedback on the task execution status. In this way, in a multi-robot system, not only is the smooth execution of the task guaranteed, but the collaborative stability of the system is also improved, and task execution interruption caused by communication delay or node failure can be avoided.

[0033] 5. The technical solution provided in this application can quickly adjust the task execution strategy and optimize the allocation of bandwidth and computing resources by sharing and feedback the real-time task status, so as to ensure that the task can be completed in a timely manner. Especially in complex and dynamic environments, it can also significantly reduce the task execution time without affecting the quality of task completion. Attached Figure Description

[0034] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0035] Figure 1 This is a schematic diagram of an embodiment of the control method for multi-machine collaborative operation based on inductive computing and control technology in this application.

[0036] Figure 2 This is a schematic diagram of an embodiment of a multi-machine collaborative operation control system based on sensing and control technology in this application. Detailed Implementation

[0037] This application provides a control method and system for multi-machine collaborative operation based on inductive computing technology. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices. Example 1

[0038] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the control method for multi-machine collaborative operation based on sensing and control technology in this application includes...

[0039] Step S1: Obtain the task granularity distribution and resource requirement matrix, and use a classification algorithm to classify the task granularity distribution and resource requirement matrix to obtain an initial task segmentation feature set; wherein obtaining the initial task segmentation feature set in step S1 includes: obtaining the task granularity distribution, resource requirement matrix and dynamic priority label; using the support vector machine algorithm to classify the task granularity distribution, resource requirement matrix and dynamic priority label to determine the classification feature vector; generating the initial task segmentation feature set based on the classification feature vector, wherein the initial task segmentation feature set includes subtask dependency weights and task granularity distribution.

[0040] Specifically, the task granularity distribution reflects the complexity and resource requirements of the task in different subtasks, while the resource requirement matrix describes the requirements of each subtask under different resource dimensions, such as computing resources, memory, bandwidth, etc. In practical applications, these data come from task analysis and resource assessment of the robot system, and are acquired in real time through sensors or scheduling systems. In step S1, the task is divided into subtasks of different granularities through sensor data collection and calculation models. The resources required for each subtask are represented in the matrix, and the priority labels of the tasks are marked to assess the urgency and importance of the tasks.

[0041] After obtaining the task granularity distribution, resource requirement matrix, and dynamic priority labels, the data is classified using the Support Vector Machine (SVM) algorithm to determine the feature vectors for task segmentation. The key to this process is leveraging the SVM's classification capabilities to categorize the task data into different classes. The SVM generates classification feature vectors by calculating the relationships between task granularity, resource requirement matrix, and priority labels. These feature vectors reflect the execution characteristics of each task. Essentially, the classification feature vectors extract classification information related to resource requirements, execution difficulty, or priority from a large amount of data. This information clarifies which subtasks have similar resource requirements, execution difficulty, or priority, thus providing a basis for subsequent task optimization.

[0042] Based on the classification feature vectors, an initial task segmentation feature set is generated, which includes subtask dependency weights and task granularity distribution. The initial task segmentation feature set not only reflects the complexity of the task but also reveals the dependencies between tasks. For example, some subtasks may need to be executed only after the preceding tasks are completed, while other subtasks can be executed in parallel. The task granularity distribution determines the task division method based on the amount of resources required and the execution time of each subtask. Through the initial task segmentation feature set, subsequent task allocation and optimization can be completed more efficiently.

[0043] This application combines classification algorithms with task allocation to efficiently handle task complexity in multi-robot systems, especially in situations where tasks are constantly changing during execution. Specifically, classification feature vectors help the system identify the urgency of tasks, resource requirements, and tasks that can be executed in parallel, thereby ensuring that task allocation conforms to the real-time resource status of the robot system. This method effectively solves the problems of unreasonable resource scheduling and untimely response in the task allocation process of multi-robot systems in the prior art, thereby improving the overall efficiency of the system and the accuracy of task execution.

[0044] Step S2: Extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and determine the optimal segmentation point set using cluster analysis. Step S2, which involves determining the optimal segmentation point set using cluster analysis, includes: extracting subtask dependency weights and task granularity distribution from the initial task segmentation feature set; clustering the subtask dependency weights using cluster analysis to obtain subtask dependency clusters; determining task segmentation points based on the subtask dependency clusters and task granularity distribution; and generating the optimal segmentation point set based on the task segmentation points, wherein the optimal segmentation point set includes subtask dependency weights and segmentation point positions.

[0045] Specifically, the task segmentation feature set is obtained from the task granularity distribution, resource requirement matrix, and dynamic priority label. It contains information such as the execution complexity, resource requirements, and priority of each subtask. Among them, the extracted subtask dependency weights reflect the dependency relationship between tasks. For example, some tasks may need to start only after other tasks are completed, while the task granularity distribution provides a fine-grained division of the computing resources required for each subtask.

[0046] After data preparation, cluster analysis is used to process the subtask dependency weights. Cluster analysis discovers the inherent relationships between tasks by grouping tasks with similar subtask dependency weights into the same category. The purpose is to assign tasks with strong dependencies to the same task group, thereby ensuring that these tasks can be executed in the appropriate order and parallelism. The clustered subtask dependency clusters show the degree of dependency between different tasks and help with subsequent task partitioning and scheduling.

[0047] Based on these dependency clusters and task granularity distributions, task splitting points can be further determined. Among these, the selection of task splitting points is crucial, as it determines how to find the optimal balance between task granularity, dependencies, and resource requirements. By analyzing the granularity distribution of tasks within dependency clusters, the most suitable splitting point location can be determined, ensuring that when tasks are assigned to different nodes, resource utilization is optimized while also preventing the disruption of task dependencies. Task splitting points are not merely simple division points; they are the result of multi-dimensional data analysis, encompassing multiple factors such as time dependencies between tasks, resource requirements, computing power, and network bandwidth.

[0048] By calculating the task split point set and combining the dependency weight and resource requirement information of each task split point, the system generates an optimal split point set. This optimal split point set provides a reasonable basis for task partitioning, ensuring that tasks can be executed efficiently in a multi-robot system. In practical applications, the selection of task split points not only affects the execution order of tasks but also the overall system's collaborative efficiency, resource allocation, and task completion timeliness. This method can effectively avoid resource conflicts, reduce task execution delays, improve task scheduling accuracy, and enhance the overall system performance.

[0049] The above method, through reasonable cluster analysis and determination of task splitting points, can solve the problems of inaccurate task allocation, resource waste, and low collaboration efficiency in existing technologies. For example, in a typical warehousing and logistics scenario, different robots perform different handling tasks. Some of these tasks require high computing resources, while others mainly rely on physical space and mobility. Without accurate task splitting and resource scheduling, the system may experience problems such as uneven resource allocation and task conflicts. Through the technical solution of this invention, the system can dynamically adjust task allocation according to task dependencies and resource requirements, avoiding resource waste, while ensuring that tasks are executed in a reasonable order and with reasonable parallelism, which can significantly improve the system's execution efficiency.

[0050] Through the above methods, the present invention can flexibly adjust task allocation and scheduling in dynamic environments, improve the collaborative efficiency and resource utilization of multi-robot systems, and solve technical problems such as delays, uneven resource allocation and poor system adaptability in task execution. It has significant technical effects and broad application prospects.

[0051] Step S3: Generate a preliminary task allocation scheme based on the optimal split point set and the node computational heterogeneity; wherein step S3, in combination with the node computational heterogeneity, generates a preliminary task allocation scheme, including: obtaining subtask dependency weights and collaborative state matrices from the optimal split point set; obtaining node computational heterogeneity, and if the node computational heterogeneity matches the subtask dependency weights, then generating a preliminary task allocation scheme based on the subtask dependency weights and collaborative state matrix, wherein the preliminary task allocation scheme includes dynamic priority labels and subtask allocation mapping.

[0052] Specifically, the optimal split point set can determine the split point location of tasks and rationally divide tasks according to resource requirements, computational load, and dependencies. Among them, the dependency weight reflects the sequential execution relationship between tasks, and the collaborative state matrix describes the collaborative needs between tasks and the collaborative capabilities between nodes. In this way, the optimal split point set can provide detailed structured information for subsequent task allocation, enabling tasks to be executed in a reasonable order while ensuring dependencies.

[0053] In the process of generating a preliminary task allocation scheme, node computational heterogeneity refers to the differences in computing power, memory resources, processing speed, etc., among different robot nodes. These differences can directly affect the efficiency and performance of task allocation. In order to allocate tasks reasonably, it is necessary to consider the computing power of each node and match it with the computing requirements of the subtasks. Specifically, the computing power of each node is evaluated in real time through the resource management module, including factors such as the node's CPU load, memory usage, processing speed, and bandwidth utilization. The node monitoring system collects data in real time and reflects it in the collaborative state matrix, which provides a comprehensive view of the current state of the nodes.

[0054] When the heterogeneity of node computation matches the dependency weight of subtasks, tasks are assigned to appropriate nodes. For example, if a subtask has a large computational requirement and depends on the completion of other tasks before it can be executed, then this task will be assigned to a node with strong computational capabilities. In this case, the execution order and computational requirements of subtasks will be optimized according to the computational capabilities of the nodes to ensure that the tasks can be executed smoothly without bottlenecks. If the computational capabilities of the nodes are insufficient, the execution progress of the tasks may be delayed, which will lead to a decrease in the collaborative efficiency of the robot system.

[0055] Based on the above matching relationship, dynamic priority tags are used to identify the priority of tasks. Priority tags are automatically generated by analyzing the urgency and importance of tasks. The priority of tasks affects the scheduling order of tasks, thereby optimizing the overall efficiency of the multi-robot system. In practical applications, tasks with higher priority will be assigned to nodes with more sufficient computing resources to ensure that tasks can be completed in a timely manner. At the same time, the generation of priority tags is also constrained by the heterogeneity of node computing, bandwidth and resources to avoid task backlog.

[0056] The subtask allocation mapping indicates which tasks are assigned to which nodes. Based on the optimal split point set, task dependency weights, node computational heterogeneity, and real-time monitoring data, an allocation mapping is generated for each task. This mapping not only ensures that tasks are executed on appropriate nodes but also adjusts the execution timing and order of tasks according to the node's resource availability, thereby optimizing task parallelism and response speed. For example, in a complex logistics warehousing scenario, multiple robots collaborate to complete the handling and sorting tasks of items. This invention can rationally allocate nodes with suitable computing power to each subtask based on task dependencies, robot node computing power, and resource requirements. Assuming a sorting task depends on the completion of a previous handling task, the order of task allocation and node allocation are automatically adjusted based on the execution progress of the former task and the robot node's computing power, ensuring that the sorting task is executed at the appropriate time and node without wasting computing resources or bandwidth.

[0057] The above methods can not only solve the problems of task execution delay, uneven distribution of computing resources and unreasonable task scheduling in traditional methods, but also improve the resource utilization efficiency and the smoothness of collaborative operation of multi-robot systems.

[0058] Step S4: Update the task granularity distribution and resource requirement matrix using real-time sensing technology to obtain a dynamically adjusted task segmentation feature set; obtaining the dynamically adjusted task segmentation feature set includes: acquiring environmental change sensitivity and cross-layer communication bandwidth; analyzing environmental change sensitivity using real-time sensing technology and updating the task granularity distribution; updating the resource requirement matrix based on cross-layer communication bandwidth; and generating the dynamically adjusted task segmentation feature set based on the updated task granularity distribution and resource requirement matrix.

[0059] Updating the task granularity distribution includes: acquiring and constructing a task granularity adjustment function based on the relationship between the current cross-layer communication bandwidth and a preset bandwidth threshold; when the current cross-layer communication bandwidth is greater than or equal to the bandwidth threshold, the task granularity adjustment function outputs a fine-grained instruction, adjusting the task granularity distribution towards finer granularity; when the current cross-layer communication bandwidth is less than the bandwidth threshold, the task granularity adjustment function outputs a granularity limiting instruction and adjusts the granularity distribution to a medium or coarse-grained distribution through a linear scaling function; based on the output of the task granularity adjustment function, dynamically updating the distribution probabilities of fine-grained, medium-grained, and coarse-grained granularities in the task granularity distribution matrix; wherein, the output of the task granularity adjustment function is also linked to the resource demand matrix.

[0060] Specifically, task granularity distribution and resource requirement matrix are important parameters describing the resource requirements and complexity among the subtasks of a task. As the task progresses, factors such as environmental changes, node status, and network bandwidth will affect the execution conditions and resource requirements of the task. Therefore, it is necessary to analyze and adjust factors such as environmental changes, node status, and network bandwidth through real-time sensing technology to maintain the efficiency of task allocation.

[0061] Environmental change sensitivity and cross-layer communication bandwidth are important factors affecting task execution efficiency. Environmental change sensitivity reflects the impact of dynamic changes during task execution on task partitioning and scheduling, while cross-layer communication bandwidth directly determines the data transmission rate and latency between nodes. Changes in these factors affect the allocation of task granularity. By monitoring and analyzing these parameters through real-time sensing technology, potential resource bottlenecks can be identified in a timely manner, and corresponding adjustments can be made. Specifically, real-time sensing technology uses sensors and data acquisition systems to acquire data on environmental change sensitivity and communication bandwidth, thereby assessing the current operating status of the system, updating the task granularity distribution and resource requirement matrix, and ensuring that tasks can be allocated and executed under appropriate time and conditions.

[0062] The task granularity adjustment function dynamically adjusts the task granularity distribution by obtaining the relationship between the current cross-layer communication bandwidth and the preset bandwidth threshold. When the cross-layer communication bandwidth is greater than or equal to the preset bandwidth threshold, the task granularity adjustment function outputs a fine-grained instruction, enabling tasks to be subdivided under higher bandwidth conditions, and the task granularity distribution will tend to be fine-grained. This fine-grained task allocation can perform more parallel processing when bandwidth is sufficient, thereby improving the system's processing power and task execution efficiency. When the cross-layer communication bandwidth is less than the bandwidth threshold, the task granularity adjustment function outputs a granularity limiting instruction, adjusting the task granularity allocation to a medium or coarse-grained distribution based on a linear proportional function. This can reduce the high bandwidth consumption caused by excessively fine task granularity and avoid communication delays and task execution bottlenecks caused by bandwidth limitations.

[0063] By dynamically updating the distribution probabilities of fine-grained, medium-grained, and coarse-grained tasks in the task granularity distribution matrix, the system can automatically adjust the task execution strategy under different bandwidth conditions, thereby adapting to different resource and environmental changes. The task granularity distribution matrix assigns different granularity levels to each subtask and adjusts them based on the current bandwidth conditions and resource requirements, thereby ensuring that tasks can be executed smoothly without increasing unnecessary computational burden. This adjustment mechanism not only optimizes the efficiency of task allocation but also maximizes the utilization of computing resources, avoiding task backlog or delays caused by bandwidth limitations.

[0064] The dynamically adjusted resource requirement matrix, supported by real-time sensing technology, reassesses the resource requirements of each subtask based on the updated task granularity distribution. By recalculating the task granularity and resource requirements, the system can allocate tasks more accurately, avoiding uneven resource allocation and overload, and ensuring that each node executes tasks efficiently within a reasonable resource range.

[0065] Through the adjustment method based on real-time perception technology described above, this invention can cope with dynamically changing task environments and complex resource requirements in multi-robot systems, improving the flexibility and response speed of task scheduling and allocation. For example, in a multi-robot collaborative warehouse management system, when a robot performs a task, the computing resources and bandwidth required depend on the complexity of the current task and the working status of other robot nodes. If the communication bandwidth in the system becomes congested, the system will automatically adjust the granularity of the task, turning fine-grained tasks into medium- or coarse-grained tasks, thereby reducing bandwidth consumption and ensuring the stability of task execution. When bandwidth is sufficient, tasks can be performed in a fine-grained manner to improve parallel processing efficiency and enhance the overall execution performance of the system.

[0066] Therefore, step S4 above, through the application of real-time perception technology, combined with task granularity adjustment and dynamic updating of the resource demand matrix, can effectively solve the problems of unreasonable task scheduling and system response lag caused by bandwidth and computing resource limitations in the existing technology, improve the task execution efficiency and resource utilization of multi-robot systems, and achieve the technical effect of optimizing multi-robot collaborative operation.

[0067] Step S5: Determine the real-time task state set through a group information sharing mechanism, and optimize the subtask dependency weights and bandwidth allocation weights using a dynamic programming algorithm to generate an optimized task allocation scheme; wherein, determining the real-time task state set through the group information sharing mechanism in step S5 includes: extracting task granularity distribution and real-time communication latency from the dynamically adjusted task segmentation feature set; distributing task execution progress, cross-layer synchronization accuracy, and real-time communication latency to each node through the group information sharing mechanism; generating a real-time task state set based on task execution progress and real-time communication latency, wherein the real-time task state set includes task execution progress and anomaly feedback vector.

[0068] Step S5 generates an optimized task allocation scheme, including: extracting task execution progress and abnormal feedback vectors from the real-time task status set; if the task execution progress is lower than a preset threshold or the abnormal feedback vector is not empty, then using a dynamic programming algorithm to optimize the subtask dependency weights and bandwidth allocation weights; and generating an optimized task allocation scheme based on the optimized subtask dependency weights and bandwidth allocation weights.

[0069] Specifically, the generation of the real-time task status set first extracts the task granularity distribution and real-time communication latency from the dynamically adjusted task segmentation feature set. This feature can reflect the complexity of the task in different subtasks, as well as the latency of data transmission between nodes. Through a group information sharing mechanism, this real-time data is distributed to each node to ensure that all nodes can obtain key information such as the current task execution progress, cross-layer synchronization accuracy, and real-time communication latency. This information can help nodes adjust their task execution strategies in a timely manner to meet the collaborative needs between different nodes and tasks.

[0070] The real-time task status set includes task execution progress and anomaly feedback vectors. The task execution progress reflects the completion status of each subtask, while the anomaly feedback vectors provide potential problems or error signals that occur during task execution, such as excessive communication latency or node failure. In multi-robot collaborative scenarios, the real-time status of tasks is crucial for system scheduling and task allocation. Therefore, the accurate generation and sharing of the real-time task status set can ensure the smooth execution of tasks and adjust the task allocation strategy based on actual feedback to avoid task delays or system crashes.

[0071] Dynamic programming algorithms optimize subtask dependency weights and bandwidth allocation weights based on task execution progress and anomaly feedback vectors. This process first assesses whether the task's execution progress meets expectations; for example, some subtasks may be progressing slowly or experiencing delays. If the task execution progress falls below a preset threshold, or if the anomaly feedback vector contains non-empty elements (e.g., excessive latency, communication interruptions), the dynamic programming algorithm adjusts the subtask dependencies and bandwidth allocation strategy to ensure the task can continue execution in the most optimal way possible. By optimizing subtask dependency weights, the algorithm can adjust the execution order between tasks, preventing overly dependent subtasks from blocking the entire task flow. Optimizing bandwidth allocation weights ensures that, with limited bandwidth, tasks can allocate bandwidth resources rationally, thereby improving data transmission efficiency and reducing latency.

[0072] After generating an optimized task allocation scheme, tasks are allocated based on the optimized subtask dependency weights and bandwidth allocation weights. The optimization scheme ensures that tasks can be executed efficiently on appropriate nodes based on the computing power and resource availability of each node. The computing resources, bandwidth, and task priority of each node are all taken into consideration, thereby maximizing the overall efficiency of the system. For example, in a logistics system where multiple robots perform tasks, multiple robots need to coordinate to move heavy objects. As the task progresses, insufficient bandwidth or some robots failing to complete the task on time due to insufficient computing power may occur. This invention can adjust the task allocation of each robot in real time and optimize the bandwidth allocation strategy to ensure that the task can be completed smoothly, solving the problems of task execution delay and resource waste caused by bandwidth limitations and node heterogeneity in the prior art.

[0073] By combining a group information sharing mechanism and a dynamic programming algorithm, this invention not only improves the flexibility and real-time performance of task allocation, but also effectively enhances the resource utilization and task execution efficiency of multi-robot systems. In particular, in complex and dynamic environments, the system can quickly respond to changes and adjust task allocation strategies, thereby ensuring the high efficiency and stability of multi-robot collaborative operations.

[0074] Step S6: Extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through a real-time communication protocol to obtain the collaborative execution result. Step S6, which updates the collaborative state matrix through a real-time communication protocol to obtain the collaborative execution result, includes: extracting task execution instructions from the optimized task allocation scheme; distributing the task execution instructions to each node through a real-time communication protocol; updating the collaborative state matrix and network latency fluctuations based on the task execution instructions; and generating the collaborative execution result based on the updated collaborative state matrix and network latency fluctuations. The collaborative execution result includes the task execution state and the node collaborative state.

[0075] Specifically, task execution instructions are extracted from the optimized task allocation scheme and distributed to each node through a real-time communication protocol. The collaborative execution results are then obtained by updating the collaborative state matrix. This process improves the overall system's collaborative efficiency and task completion rate by ensuring the smooth execution and collaborative cooperation of tasks among multiple nodes.

[0076] Specifically, the optimized task allocation scheme is generated in step S5 above. This scheme comprehensively considers the task execution progress, subtask dependencies, and resource allocation, and optimizes the task execution order and bandwidth allocation through a dynamic programming algorithm. Based on this scheme, the system can determine the execution priority of each subtask, which node it is assigned to, and the resource requirements of that task. This information constitutes the task execution instructions, which reflect the resources required by each task or subtask and its execution time window.

[0077] After receiving the task execution instructions, the system transmits these instructions to each node through a real-time communication protocol. The communication protocol ensures that the instructions can reach each execution node quickly and accurately, and guarantees that the execution order of the tasks is synchronized with the time. In order to cope with the impact of factors such as network latency and bandwidth fluctuations, the real-time communication protocol also involves a mechanism for confirming and retransmitting feedback information from each node, ensuring that all instructions are delivered to each node in a timely and effective manner.

[0078] After the task execution instructions are transmitted via a real-time communication protocol, the collaborative state matrix is ​​updated to reflect the task execution status and the collaboration between nodes. The collaborative state matrix is ​​a real-time dynamic data structure that describes the current execution status, resource usage, and task progress of each node. It includes not only the execution progress of each subtask but also the collaboration efficiency, communication latency, and bandwidth utilization between nodes. Based on the collaborative state matrix, dynamic adjustments can be made during task execution to ensure that the task can be completed smoothly according to the predetermined plan.

[0079] By updating the coordination state matrix, fluctuations in network latency can be reflected, and task execution can be corrected in a timely manner. For example, if a node fails to receive a task instruction on time due to network latency, the coordination state matrix will record this latency and trigger a fast recovery mechanism to ensure that the node synchronizes as soon as possible and continues task execution. This mechanism can respond to latency and error feedback in a timely manner, ensuring the smooth execution of the overall task.

[0080] In multi-robot systems, collaborative execution results not only reflect task completion status but also the collaborative efficiency between nodes, directly impacting overall operational efficiency and resource utilization. In a real-world logistics task, if a node's task execution is hindered, the updated collaborative state matrix can identify that node's status and adjust resource or task allocation, preventing other nodes from wasting time or resources waiting. Therefore, collaborative execution results not only reflect task completion but also provide a basis for optimization and adjustment, enabling multi-robot systems to efficiently and stably complete complex tasks. For example, in robotic warehouse management scenarios, multiple robots need to collaboratively move goods. During execution, if a robot experiences data transmission delays due to bandwidth limitations, the real-time communication protocol will reallocate bandwidth and update the collaborative state matrix through a feedback mechanism, allowing the task execution order to be adjusted based on the current network status. By analyzing collaborative execution results, the system can promptly identify problems and take corrective measures, ensuring that the overall system's collaborative efficiency remains unaffected.

[0081] By updating the collaborative state matrix through a real-time communication protocol, this invention effectively solves the problems of task execution delay, resource conflict, and unstable node collaboration in the prior art. This technical means can reflect the status of task execution in real time, dynamically adjust the task allocation scheme, and improve the adaptability and execution efficiency of multi-robot systems. Especially in complex and ever-changing working environments, this technical solution demonstrates significant technical advantages.

[0082] Through the coordination of the above steps, the present invention can dynamically adjust task allocation, optimize resource utilization, and improve the collaborative efficiency of multi-robot systems in complex environments. Example 2

[0083] The control method for multi-machine collaborative operation based on sensing and computing technology in the embodiments of this application has been described above. The control system for multi-machine collaborative operation based on sensing and computing technology in the embodiments of this application is described below. Please refer to [link to relevant documentation]. Figure 2 One embodiment of the multi-machine collaborative operation control system based on sensing and control technology in this application includes...

[0084] The initial task segmentation unit is used to obtain the task granularity distribution and resource requirement matrix. A classification algorithm is used to classify the task granularity distribution and resource requirement matrix to obtain the initial task segmentation feature set.

[0085] The optimal segmentation unit is determined to extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and cluster analysis is used to determine the optimal segmentation point set.

[0086] The preliminary task allocation unit is used to generate a preliminary task allocation scheme based on the optimal set of split points and the heterogeneity of node computation.

[0087] The dynamic task segmentation unit is used to update the task granularity distribution and resource requirement matrix through real-time sensing technology to obtain a dynamically adjusted task segmentation feature set.

[0088] The task allocation unit is optimized to determine the real-time task status set through a group information sharing mechanism, and to optimize the subtask dependency weights and bandwidth allocation weights using a dynamic programming algorithm to generate an optimized task allocation scheme.

[0089] The collaborative execution unit is used to extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through a real-time communication protocol to obtain the collaborative execution results.

[0090] Through the synergistic cooperation of the above-mentioned components, the collaborative efficiency of the multi-robot system is further improved.

[0091] It should also be noted that the various specific technical features described in the above embodiments can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, this application will not describe the various possible combinations separately. Furthermore, various different embodiments of this application can also be arbitrarily combined, as long as they do not violate the spirit of this invention, and should also be regarded as the content disclosed by this invention.

[0092] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0093] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0094] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A control method for multi-machine collaborative operation based on sensing and control technology, characterized in that, The method includes: Step S1: Obtain the task granularity distribution and resource requirement matrix, and use a classification algorithm to classify the task granularity distribution and resource requirement matrix to obtain the initial task segmentation feature set; Step S2: Extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and use cluster analysis to determine the optimal segmentation point set; Step S3: Based on the optimal split point set, and considering the heterogeneity of node computation, generate a preliminary task allocation scheme; generating a preliminary task allocation scheme by considering the heterogeneity of node computation includes: Obtain the subtask dependency weights and collaborative state matrix from the optimal split point set; Obtain the node computation heterogeneity. If the node computation heterogeneity matches the subtask dependency weight, generate a preliminary task allocation scheme based on the subtask dependency weight and the cooperative state matrix. The preliminary task allocation scheme includes dynamic priority labels and subtask allocation mapping. Step S4: Update the task granularity distribution and resource requirement matrix using real-time sensing technology to obtain a dynamically adjusted task segmentation feature set; the dynamically adjusted task segmentation feature set includes: Acquire sensitivity to environmental changes and cross-layer communication bandwidth; By analyzing the sensitivity to environmental changes through real-time sensing technology, the task granularity distribution is updated. Update resource demand matrix based on cross-layer communication bandwidth; Based on the updated task granularity distribution and resource requirement matrix, a dynamically adjusted task segmentation feature set is generated. Update task granularity distribution, including: Obtain and construct a task granularity adjustment function based on the relationship between the current cross-layer communication bandwidth and the preset bandwidth threshold; When the current cross-layer communication bandwidth is greater than or equal to the bandwidth threshold, the task granularity adjustment function outputs fine-grained instructions to adjust the task granularity distribution towards fine-grainedness. When the current cross-layer communication bandwidth is less than the bandwidth threshold, the task granularity adjustment function outputs a granularity limit instruction and adjusts the granularity distribution to a medium or coarse granular distribution through a linear scaling function. Based on the output of the task granularity adjustment function, the distribution probabilities of fine-grained, medium-grained, and coarse-grained tasks in the task granularity distribution matrix are dynamically updated; the output of the task granularity adjustment function is also linked to the resource requirement matrix. Step S5: Determine the real-time task state set through the group information sharing mechanism, and use the dynamic programming algorithm to optimize the subtask dependency weights and bandwidth allocation weights to generate an optimized task allocation scheme; Step S6: Extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through the real-time communication protocol to obtain the collaborative execution results.

2. The method according to claim 1, characterized in that, The initial task segmentation feature set obtained in step S1 includes: Obtain task granularity distribution, resource requirement matrix, and dynamic priority tags; The support vector machine algorithm is used to classify the task granularity distribution, resource requirement matrix, and dynamic priority labels to determine the classification feature vector; An initial task segmentation feature set is generated based on the classification feature vector, wherein the initial task segmentation feature set includes subtask dependency weights and task granularity distribution.

3. The method according to claim 1, characterized in that, Step S2 uses cluster analysis to determine the optimal split point set, including: Extract subtask dependency weights and task granularity distribution from the initial task segmentation feature set; Cluster analysis techniques are used to cluster the subtask dependency weights to obtain subtask dependency clusters; The task splitting point is determined based on the subtask dependency clusters and task granularity distribution; An optimal split point set is generated based on the task split points, where the optimal split point set includes the subtask dependency weights and the split point positions.

4. The method according to claim 1, characterized in that, Step S5, which determines the real-time task status set through a group information sharing mechanism, includes: Extract task granularity distribution and real-time communication latency from the dynamically adjusted task segmentation feature set; The task execution progress, cross-layer synchronization accuracy, and real-time communication latency are distributed to each node through a group information sharing mechanism. Based on task execution progress and real-time communication latency, a real-time task status set is generated, which includes task execution progress and anomaly feedback vectors.

5. The method according to claim 1, characterized in that, The step S5, generating an optimized task allocation scheme, includes: Extract task execution progress and anomaly feedback vectors from the real-time task status set; If the task execution progress is lower than the preset threshold or the abnormal feedback vector is not empty, a dynamic programming algorithm is used to optimize the subtask dependency weights and bandwidth allocation weights. An optimized task allocation scheme is generated based on the optimized subtask dependency weights and bandwidth allocation weights.

6. The method according to claim 1, characterized in that, In step S6, the collaborative state matrix is ​​updated through a real-time communication protocol to obtain the collaborative execution result, including: Extract task execution instructions from the optimized task allocation scheme; The task execution instructions are distributed to each node via a real-time communication protocol; Update the collaborative state matrix and network latency fluctuations based on task execution instructions; Based on the updated collaborative state matrix and network latency fluctuations, collaborative execution results are generated, which include task execution status and node collaborative status.

7. A control system for multi-machine collaborative operation based on sensing and computing technology, used to implement the control method for multi-machine collaborative operation based on sensing and computing technology as described in any one of claims 1-6, characterized in that, The system includes: The initial task segmentation unit is used to obtain the task granularity distribution and resource requirement matrix. A classification algorithm is used to classify the task granularity distribution and resource requirement matrix to obtain the initial task segmentation feature set. The optimal segmentation unit is determined to extract subtask dependency weights and task granularity distribution based on the initial task segmentation feature set, and cluster analysis is used to determine the optimal segmentation point set. The preliminary task allocation unit is used to generate a preliminary task allocation scheme based on the optimal set of split points and the heterogeneity of node computation. The dynamic task segmentation unit is used to update the task granularity distribution and resource requirement matrix through real-time sensing technology to obtain a dynamically adjusted task segmentation feature set. The task allocation unit is optimized to determine the real-time task status set through a group information sharing mechanism, and to optimize the subtask dependency weights and bandwidth allocation weights using a dynamic programming algorithm to generate an optimized task allocation scheme. The collaborative execution unit is used to extract task execution instructions from the optimized task allocation scheme, distribute them to each node, and update the collaborative state matrix through a real-time communication protocol to obtain the collaborative execution results.

Images