AIoT cloud edge-end intelligent control method and system for industrial internet of things

Abstract

The application provides an AIoT cloud edge end intelligent control method and system for an industrial Internet of Things, relates to the technical field of cloud edge end intelligent control, and comprises the following steps: creating a cloud-edge-end collaborative architecture; performing adaptive analysis on intelligent manufacturing tasks according to device performance and device functions, establishing a first adaptive constraint; performing task decomposition of the intelligent manufacturing tasks by using a dynamic task decomposition channel, and creating M subtasks; performing task competition game analysis of the terminal devices on the M subtasks by using a cloud center, establishing a second adaptive constraint; and performing balance analysis on the first adaptive constraint and the second adaptive constraint, and generating an intelligent execution scheme of the terminal devices for the intelligent manufacturing tasks. The application can achieve the technical goal of intelligent task scheduling and optimized allocation under the cloud-edge-end collaborative computing architecture, improve task execution efficiency, reduce computing delay, optimize computing resource utilization, and enhance data security.

Description

Technical Field

[0001] This application relates to the field of cloud-edge-device intelligent control technology, and in particular to AIoT cloud-edge-device intelligent control methods and systems for the industrial Internet of Things. Background Technology

[0002] In the context of the development of smart manufacturing and the Industrial Internet of Things (IIoT), how to efficiently allocate and execute complex computing tasks has become a key issue. Traditional task scheduling methods typically rely on cloud computing centers for unified management and allocation. However, this model has certain limitations, especially when facing the complex and ever-changing computing demands in the IIoT environment. A single cloud-based processing approach often leads to uneven distribution of computing resources, low task execution efficiency, and slow system response. Therefore, cloud-edge-device collaborative computing architecture has gradually become an important technological direction for solving this problem. It combines the powerful computing capabilities of the cloud, the real-time processing capabilities of edge computing, and the distributed execution capabilities of terminal devices to optimize task scheduling and improve computing efficiency.

[0003] Currently, existing cloud computing models primarily rely on centralized computing architectures, where all computing tasks must be uploaded to a remote cloud for processing before the results are returned to the terminal devices. While this model offers powerful computing capabilities, it also has several significant drawbacks. First, because cloud computing requires data transmission over a network, traditional cloud computing models often struggle to meet the high real-time requirements of tasks in industrial IoT scenarios. For example, in smart manufacturing systems, certain critical tasks, such as equipment fault detection and production process optimization, need to complete calculations and decisions within a very short timeframe. The high latency of traditional cloud computing can lead to untimely system responses, thus impacting production efficiency. Second, as the number of industrial IoT devices continues to increase, the data transmission load also rises. Traditional cloud computing models are easily limited by network bandwidth, leading to data transmission congestion and affecting task execution efficiency. Furthermore, a single cloud computing model may also face data security and privacy issues. For instance, uploading all sensitive production data to the cloud for processing could increase the risk of data leakage.

[0004] In summary, existing technologies suffer from the technical problems caused by the reliance on centralized computing architecture in cloud computing models, which necessitates remote transmission of computing tasks. This leads to high latency, limited network bandwidth, and data security risks, further impacting the demands for real-time performance, efficiency, and data privacy in industrial IoT scenarios, and ultimately reducing production efficiency and system stability. Summary of the Invention

[0005] The purpose of this application is to provide an AIoT cloud-edge-device intelligent control method and system for the Industrial Internet of Things (IIoT), in order to solve the technical problems in the prior art where the cloud computing model relies on a centralized computing architecture, which requires computing tasks to be transmitted remotely, resulting in high latency, limited network bandwidth, and data security risks. This further affects the requirements for real-time performance, efficiency, and data privacy in the IIoT scenario, thereby reducing production efficiency and system stability.

[0006] In view of the above problems, this application provides an AIoT cloud-edge-device intelligent control method and system for the Industrial Internet of Things.

[0007] Firstly, this application provides an AIoT cloud-edge-device intelligent control method for the Industrial Internet of Things (IIoT), implemented through an AIoT cloud-edge-device intelligent control system for the IIoT. The method includes: creating a cloud-edge-device collaborative architecture, comprising a cloud center, edge computing nodes, and terminal devices; obtaining intelligent manufacturing tasks from the cloud center; acquiring the device performance and functions of the terminal devices at the edge computing nodes; performing an adaptation analysis of the intelligent manufacturing tasks based on the device performance and functions to establish a first adaptation constraint; activating a dynamic task decomposition channel in the cloud center; using the dynamic task decomposition channel to decompose the intelligent manufacturing tasks into M sub-tasks; using the cloud center to perform a task competition game analysis of the terminal devices on the M sub-tasks to establish a second adaptation constraint; and performing a balance analysis on the first and second adaptation constraints to generate an intelligent execution scheme for the terminal devices of the intelligent manufacturing tasks.

[0008] Secondly, this application also provides an AIoT cloud-edge-device intelligent control system for the Industrial Internet of Things (IIoT), used to execute the AIoT cloud-edge-device intelligent control method for the Industrial Internet of Things as described in the first aspect, comprising: a collaborative architecture creation module for creating a cloud-edge-device collaborative architecture, the cloud-edge-device collaborative architecture including a cloud center, edge computing nodes, and terminal devices, and obtaining intelligent manufacturing tasks from the cloud center; an adaptation analysis module for obtaining the device performance and device functions of the terminal devices at the edge computing nodes, performing adaptation analysis on the intelligent manufacturing tasks based on the device performance and device functions, and establishing a first adaptation constraint; a task decomposition module for activating the dynamic task decomposition channel of the cloud center, using the dynamic task decomposition channel to decompose the intelligent manufacturing tasks, and creating M sub-tasks; a competitive game analysis module for using the cloud center to perform task competitive game analysis on the terminal devices for the M sub-tasks, and establishing a second adaptation constraint; and a scheme generation module for performing balance analysis on the first adaptation constraint and the second adaptation constraint, and generating an intelligent execution scheme for the terminal devices of the intelligent manufacturing tasks.

[0009] The technical solution provided in this application has at least the following technical effects or advantages: by realizing the technical goal of intelligent task scheduling and optimized allocation under the cloud-edge-device collaborative computing architecture, it achieves the technical effects of improving task execution efficiency, reducing computing latency, optimizing computing resource utilization, and enhancing data security.

[0010] The above description is merely an overview of the technical solution of this application. To better understand the technical means of this application and to facilitate its implementation according to the description, and to make the above and other objects, features, and advantages of this application more apparent, specific embodiments of this application are described below. It should be understood that the content described in this section is not intended to identify key or important features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent through the following description. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are merely exemplary. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0012] Figure 1 This is a flowchart illustrating the AIoT cloud-edge-device intelligent control method for the industrial Internet of Things (AIIoT) proposed in this application.

[0013] Figure 2 This is a schematic diagram of the structure of the AIoT cloud-edge-device intelligent control system for industrial IoT in this application.

[0014] Figure labeling: Collaborative architecture creation module 11, adaptation analysis module 12, task decomposition module 13, competitive game analysis module 14, solution generation module 15. Detailed Implementation

[0015] This application provides an AIoT cloud-edge-device intelligent control method and system for the Industrial Internet of Things (IIoT). It addresses the technical problems in existing technologies where cloud computing relies on centralized computing architectures, leading to high latency, limited network bandwidth, and data security risks due to the need for remote transmission of computing tasks. These issues further impact the real-time performance, efficiency, and data privacy requirements of IIoT scenarios, ultimately reducing production efficiency and system stability. The application achieves the technical goal of intelligent task scheduling and optimized allocation under a cloud-edge-device collaborative computing architecture, resulting in improved task execution efficiency, reduced computational latency, optimized computing resource utilization, and enhanced data security.

[0016] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. It should be understood that this application is not limited to the exemplary embodiments described herein. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application. It should also be noted that, for ease of description, only the parts related to this application are shown in the accompanying drawings, not all of them.

[0017] Example 1, please refer to the appendix. Figure 1 This application provides an AIoT cloud-edge-device intelligent control method for industrial IoT, which is applied to an AIoT cloud-edge-device intelligent control system for industrial IoT, and specifically includes the following steps:

[0018] S1: Create a cloud-edge-device collaborative architecture, which includes a cloud center, edge computing nodes, and terminal devices, and obtain intelligent manufacturing tasks from the cloud center.

[0019] Specifically, a cloud-edge-device collaborative architecture is created, building a system architecture consisting of three parts: the cloud, edge computing, and terminal devices. The cloud center is responsible for processing and storing large amounts of data. Based on real-time collected environmental and device data, relevant task instructions are generated and then distributed to edge computing nodes or terminal devices. Edge computing nodes are responsible for data processing and analysis closer to the terminal devices, while the terminal devices are responsible for interacting with the real-world environment and executing actual tasks. Through the collaborative work of the cloud, edge, and device, smooth data transmission and efficient execution of intelligent decisions can be achieved.

[0020] S2: Obtain the device performance and device functions of the terminal device at the edge computing node, perform adaptation analysis of the intelligent manufacturing task based on the device performance and device functions, and establish the first adaptation constraint.

[0021] Specifically, edge computing nodes acquire the device performance and device functions of terminal devices. Device performance refers to the hardware indicators of terminal devices, such as processing power, computing speed, and memory, while device functions refer to the specific tasks that terminal devices can perform, such as detection, monitoring, and control.

[0022] Next, an adaptation analysis of the intelligent manufacturing task is conducted based on the equipment performance and functions to assess whether the terminal equipment can handle the current intelligent manufacturing task. Intelligent manufacturing tasks include automated control, data acquisition, and process optimization. Based on the analysis results, certain constraints are set for the matching between equipment performance and functions and the intelligent manufacturing task, establishing the first adaptation constraint. This first adaptation constraint ensures that the equipment's performance and functions meet the task requirements, avoiding resource waste and task execution failure.

[0023] S3: Activate the dynamic task decomposition channel in the cloud center, and use the dynamic task decomposition channel to decompose the intelligent manufacturing task and create M sub-tasks.

[0024] Specifically, this involves activating the dynamic task decomposition channel in the cloud center. The dynamic task decomposition channel is a mechanism for task splitting and scheduling, capable of flexibly adjusting task decomposition strategies based on the system's real-time computing needs and device status. For example, in some situations, a task may need to be subdivided into multiple smaller tasks for parallel processing, while in situations with limited computing resources, the number of task splits may need to be reduced to accommodate computing power limitations. Therefore, activating the dynamic task decomposition channel in the cloud center can make task decomposition more flexible and intelligent, improving overall computing efficiency.

[0025] Next, the task decomposition of intelligent manufacturing tasks using a dynamic task decomposition channel means that, with the support of the cloud center, the intelligent manufacturing task is broken down into smaller sub-tasks through this channel. The task decomposition process needs to consider multiple factors, such as the task's computational requirements, the computing power of the equipment, and the stability of network communication. For example, a large machine vision inspection task might be decomposed into multiple sub-tasks, each responsible for a different detection area or different feature extraction operations. This allows multiple computing devices to process the data in parallel, improving inspection efficiency. Furthermore, the dynamic task decomposition channel can adjust the task decomposition strategy based on real-time computational load. For example, with sufficient computing resources, a more refined task split can be used to increase computational parallelism; while with limited resources, task splitting can be reduced to lower communication and scheduling costs.

[0026] Finally, after task decomposition, M executable subtasks are generated. The value of M depends on the complexity of the task, the availability of computing resources, and the execution method. For example, if a smart manufacturing task involves complex data processing and computing resources are sufficient, M may be large, such as 50 subtasks, each executed on a different computing device; while if computing resources are limited, M may be small, such as only 10 subtasks, to reduce competition for computing resources. Reasonable task decomposition can improve computational efficiency, reduce waste of computing resources, and ensure that the task can be completed within the specified time.

[0027] S4: Utilize the cloud center to perform task competition game analysis on the terminal devices for M sub-tasks, and establish a second adaptation constraint.

[0028] Specifically, the M subtasks are multiple executable tasks created through a dynamic task decomposition channel, each potentially with different computational requirements and execution priorities. Task competition game analysis is a game theory-based computational method designed to solve the problem of multiple tasks competing for limited computing resources. For example, when multiple subtasks need to be assigned to different terminal devices, each device's computing power, current load, and task suitability will affect the task scheduling outcome. Through task competition game analysis, an optimal task allocation strategy can be found that maximizes task execution efficiency while simultaneously maximizing device resource utilization.

[0029] Next, after the task competition game analysis is completed, more reasonable task allocation constraints are formulated based on the calculation results as the second adaptation constraint to ensure that tasks are reasonably allocated to the most suitable terminal devices. If a terminal device has low computing power, the task may be avoided from being assigned to that device, thereby reducing the possibility of computational failure or execution delay. The second adaptation constraint comprehensively considers the results of the task competition game analysis, making task allocation more intelligent and efficient. For example, in the first adaptation constraint, a task may be considered suitable for execution on multiple devices, while in the second adaptation constraint, the optimal device is ultimately selected to execute the task, taking into account the task competition situation and the real-time computing load of the devices.

[0030] S5: Perform a balance analysis on the first adaptation constraint and the second adaptation constraint to generate an intelligent execution plan for the terminal equipment of the intelligent manufacturing task.

[0031] Specifically, a comprehensive comparison of the first and second adaptation constraints is conducted to ensure the rationality of task allocation. The goal of the balance analysis is to find the optimal compromise between the two, so that tasks can be matched with suitable equipment while also being dynamically optimized based on real-time computing status. For example, if the first adaptation constraint determines that a certain device is suitable for executing high-computation tasks, but the second adaptation constraint finds that the device is currently overloaded, the task may be assigned to a device with a lower load but slightly lower computing power to improve overall execution efficiency. Balance analysis typically involves multiple optimization parameters, such as computing resource utilization, task execution time, communication stability, and task completion rate. Finally, based on the results of the balance analysis, an intelligent execution scheme for the terminal equipment of intelligent manufacturing tasks is generated. The intelligent execution scheme is an optimized task scheduling scheme that can rationally allocate tasks based on the equipment's computing power, task requirements, real-time load, and historical execution performance.

[0032] Furthermore, this application also includes: activating the task competition game analysis function of the cloud center, performing task competition game analysis of M sub-tasks, wherein the game characteristics of the competition game analysis function include device adaptability characteristics, historical performance consistency characteristics, communication stability characteristics, and data fidelity guarantee characteristics, and the task competition game analysis function performs game compensation through inter-task collaboration factors.

[0033] Specifically, the task competition game analysis function is as follows: ;in, Characterization subtask In terminal device resources Losses in competitive games Representation task Dynamic priority weights, Equipment adaptability factors, or equipment adaptability characteristics, are used to measure the resources of terminal equipment. Pair Task Adaptability Characterizing terminal device resources Historical performance consistency, i.e., the characteristic of historical performance consistency. Factors characterizing communication stability, i.e., communication stability features. The data fidelity guarantee factor, i.e., the data fidelity guarantee characteristic, Characterize the synergistic factors between tasks. These are the weighting factors for equipment adaptability, historical performance consistency, communication stability, and data fidelity assurance, respectively. This represents the loss weighting factor for inter-task collaboration. The task competition game analysis function in the cloud center is then initiated. Task competition game analysis is a game theory-based method used to solve the problem of multiple tasks competing for limited resources. The analysis function is used to calculate and evaluate the competitive relationships between tasks in order to optimize task allocation strategies.

[0034] Then, a game theory analysis of the task competition among M subtasks is performed to determine how these subtasks can rationally utilize terminal device resources and avoid excessive resource contention leading to system performance degradation. M represents the number of tasks to be executed. A subtask is a smaller-scale computation or operation resulting from the breakdown of a complete task; each subtask requires a certain amount of computing resources to complete.

[0035] Next, Characterization subtask In terminal device resources The competitive game loss is used to measure the efficiency loss caused by multiple subtasks competing for terminal device resources. When multiple tasks compete for the same computing resources simultaneously, delays, task failures, or decreased computing resource utilization may occur.

[0036] Representation task Dynamic priority weighting refers to a variable used to measure the priority of different tasks. Dynamic priority means that the priority of tasks adjusts as time or circumstances change. For example, the priority of urgent tasks may increase, while the priority of less important tasks may be postponed to ensure the rational allocation of resources.

[0037] at the same time, The device adaptability factor measures the ability of a terminal device's resources to adapt to sub-tasks, representing the device's ability to adapt to different task requirements. For example, a high-performance computing device may be able to handle multiple complex tasks simultaneously, while a low-power device may only be able to perform simple calculations. The higher this factor, the stronger the device's ability to adapt to different tasks.

[0038] Characterizing terminal device resources Historical performance consistency refers to a variable used to measure whether the performance of a terminal device has been stable over the past. If the device's computing speed, response time, and other parameters fluctuate little over different time periods, it indicates high historical performance consistency, meaning it can execute tasks more reliably.

[0039] The communication stability factor measures the quality of communication between a terminal device and the cloud or other devices. A high communication stability factor indicates a stable network connection, low data transmission latency, and a lower likelihood of data loss or transmission interruption, thus ensuring the normal execution of tasks.

[0040] The data fidelity factor is a parameter used to measure whether data can maintain its integrity and accuracy during transmission and processing. A higher data fidelity factor means that the data is less affected by interference during transmission, has lower errors, and the final calculation results are more reliable.

[0041] The inter-task collaboration factor represents the degree to which multiple tasks work together. A high collaboration factor indicates that different tasks can efficiently share resources, reduce resource conflicts, and improve overall system efficiency. For example, on a manufacturing production line, different tasks can share sensor data to optimize the entire production process.

[0042] These are weighting factors for device adaptability, historical performance consistency, communication stability, and data fidelity assurance, respectively. These factors adjust the weight of each parameter's impact on the final calculation. Different application scenarios may require different weights. For example, in systems with high real-time requirements, the weighting factor for communication stability may be larger, while in data analysis tasks, the weighting factor for data fidelity may be higher. Additionally, there are weighting factors for inter-task collaboration. This is used to measure the impact of the degree of collaboration between tasks on the overall loss. If the task collaboration capability is insufficient, it will lead to a decrease in resource utilization efficiency. Therefore, it is necessary to set an appropriate loss weight factor for optimization.

[0043] The inter-task synergy factor is calculated as follows: ;in, Characterized in terminal device resources The total number of subtasks scheduled above. Characterization subtask With another sub-task The collaborative similarity, and Belongs to terminal device resources The set of subtasks scheduled above.

[0044] First, the inter-task coordination factor is a parameter used to measure the collaborative work of multiple subtasks when executing on terminal device resources. It reflects the degree of mutual influence between tasks. Calculating the inter-task coordination factor is to optimize task scheduling, enabling multiple tasks to efficiently share resources, reduce resource waste, and improve overall execution efficiency.

[0045] in, Characterized in terminal device resources The total number of scheduled subtasks refers to the number of subtasks running simultaneously on a specific terminal device. Terminal device resources include limited resources such as computing power, storage space, and network bandwidth. When multiple subtasks run simultaneously, they need to be scheduled reasonably to ensure maximum resource utilization. For example, if a terminal device can process five subtasks simultaneously, but only three subtasks are actually running, it means that the resources are not being fully utilized. If the number of running subtasks exceeds its processing capacity, it may cause computational delays or failures.

[0046] Characterization subtask With another sub-task The co-operational similarity variable measures how closely two subtasks work together during execution. High co-operational similarity means that two subtasks can share computational results or data inputs. For example, in image processing tasks, if one subtask is responsible for image segmentation and the other for object detection, they can share preprocessed data, reducing redundant computations and improving execution efficiency. Conversely, if two subtasks require independent data processing, low co-operational similarity indicates less resource sharing and may increase computational costs.

[0047] Belongs to terminal device resources The set of subtasks scheduled means that the subtasks participating in the calculation of inter-task coordination factors must be part of the set of tasks currently running on the terminal device. The set of subtasks refers to all tasks scheduled to run on a particular terminal device, not all tasks in the entire system. For example, if a terminal device is running four subtasks, and the entire system has ten subtasks, then only these four tasks need to be considered when calculating the coordination factor, not all ten. This constraint ensures the specificity of the calculation, making task scheduling more accurate.

[0048] Furthermore, this application also includes: evaluating the computing power of the terminal device based on its device performance, and establishing a weak computing device performance protection identifier; performing load analysis on the terminal device within a preset period, and establishing a high load identifier; using the weak computing device performance protection identifier and the high load identifier to perform equilibrium compensation for task competition game analysis, and establishing a second adaptation constraint based on the equilibrium compensation result.

[0049] Specifically, computing power evaluation refers to assessing the task processing capabilities of terminal devices based on their performance parameters in order to allocate computing tasks appropriately. For example, if a device has strong computing power, it can be assigned complex tasks, while devices with weaker computing power can only perform simple tasks, thereby improving the overall system efficiency.

[0050] Next, a performance protection flag is established for low-computing devices. This means that if certain terminal devices have low computing power, a special flag will be assigned to them to prevent them from undertaking excessively heavy computing tasks. Low-computing devices typically refer to devices with low processor performance, limited memory, or power-constrained devices, such as small sensor nodes or low-power embedded devices. The purpose of the protection flag is to limit the task allocation of these devices, ensuring that they do not experience performance degradation or even crash due to overloading.

[0051] Then, the operating load of the terminal devices is periodically assessed according to the set time intervals. The preset period range refers to a time window set by those skilled in the art, such as performing load checks every ten minutes or every hour. Load analysis mainly focuses on indicators such as CPU utilization, memory usage, and data throughput of the devices to determine their current operating status. If a device is under high load for an extended period, it may affect the stability of the system, so measures need to be taken to optimize scheduling.

[0052] Next, a high-load flag is established. This means that if a terminal device is detected to have a consistently high load during load analysis, a high-load flag is assigned to it to alert the task scheduling system and prevent further computing tasks from being assigned to that device, thus preventing resource exhaustion or task execution failure. For example, if a device's CPU utilization remains at 90% for an extended period, it may be unable to handle new computing tasks, and the high-load flag will mark it as busy, thereby optimizing the task allocation strategy.

[0053] Furthermore, utilizing the performance protection flags and high-load flags of weak computing devices for equilibrium compensation in task competition game analysis implies comprehensively considering the computing power and current load of the devices to optimize the task allocation scheme. Task competition game analysis is a computational method based on game theory, aiming to solve the problem of multiple tasks competing for limited computing resources. Equilibrium compensation refers to ensuring that the computing resources of different devices are optimally utilized by reasonably adjusting the task allocation, preventing both overloading of devices with weak computing power and over-capacity operation of devices with high load. For example, if a device with weak computing power originally needs to execute five tasks, but due to insufficient computing power, the allocation scheme can be adjusted so that a device with stronger computing power can take over two of the tasks, thereby achieving equilibrium compensation.

[0054] Finally, a second adaptation constraint is established based on the equilibrium compensation results. This means that after the task competition game analysis is completed, the matching relationship between tasks and devices is further optimized based on the equilibrium compensation results. The role of the second adaptation constraint is to ensure that task allocation is more reasonable and to avoid resource waste or computational bottlenecks. For example, if a device was considered suitable for performing a certain task in the previous adaptation analysis, but its adaptability changes due to subsequent load increases, then the second adaptation constraint will re-evaluate the task allocation and adjust the resource usage strategy to maintain the stability and efficiency of the system.

[0055] Furthermore, this application also includes: uploading the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints of the intelligent manufacturing task; obtaining the highest and lowest computing power of the terminal device, and establishing granularity decomposition constraints based on the highest and lowest computing power; and decomposing the intelligent manufacturing task into M sub-tasks based on the granularity decomposition constraints, the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints.

[0056] Specifically, the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints of intelligent manufacturing tasks are uploaded for subsequent analysis and processing. Intelligent manufacturing tasks refer to the calculations or operations that need to be performed during industrial production, such as production scheduling, equipment control, and data analysis. Time constraints mean that the task must be completed within a specified time; for example, the processing of a certain part must be completed within 20 minutes, otherwise it will affect the entire production process. Computational complexity constraints are a standard for measuring the computational resources required for a task; they determine how much computing power is needed to complete the task. For example, complex artificial intelligence model inference requires high computational complexity, while simple temperature monitoring tasks have lower computational complexity. Computational requirement constraints refer to the computational resources required by the task, including CPU, memory, and storage. For example, one task may require high-performance GPU computing, while another task only requires basic numerical operations. Task association constraints refer to the interdependencies between different tasks; for example, the output of one task may be the input of another task, therefore they must be executed in a certain order.

[0057] Next, obtaining the highest and lowest computing power of the terminal devices means testing the computing power of all available terminal devices. The highest computing power refers to the capability of the most powerful device, such as a high-performance computing server or an industrial-grade edge computing node. The lowest computing power refers to the device with the weakest computing power in the system, such as a low-power sensor or an embedded device. Obtaining these two extreme values ​​aims to assess the overall computing power range of the system and provide a basis for subsequent task decomposition.

[0058] Then, granularity decomposition constraints are established based on the highest and lowest computing power. This means setting the task decomposition granularity according to the range of device computing capabilities. Granularity decomposition constraints determine how many subtasks a task is broken down into and the computational load of each subtask. If the computing power range is large, indicating the availability of powerful computing devices, the task can be broken down into finer granularities, allowing multiple devices to process in parallel and improving overall efficiency. For example, if a task requires 100 units of computation, and the device with the highest computing power can process 50 units per second, while the device with the lowest computing power can only process 5 units per second, then the system might break the task down into multiple subtasks of different sizes and assign them to devices with different computing capabilities to achieve optimal computation scheduling.

[0059] Finally, based on granularity decomposition constraints, time constraints, computational complexity constraints, computational requirement constraints, and task association constraints, the intelligent manufacturing task is decomposed into M sub-tasks to create more rationally allocated execution across different terminal devices. Task decomposition is a crucial method for optimizing the utilization of computing resources in intelligent manufacturing systems. For example, a large data analysis task may be broken down into multiple smaller, parallel computing tasks, each executed by different devices, thereby accelerating computation. If time constraints are tight, tasks may be prioritized for devices with high computing power; if computational requirements are high, task allocation to devices with low computing power may be minimized to ensure successful task completion. Table 1 shows the most recent intelligent manufacturing task and terminal device computing power data record.

[0060] Table 1: Data Record of the Most Recent Intelligent Manufacturing Task and Terminal Equipment Computing Capability

[0061]

[0062] Furthermore, this application also includes: monitoring the execution of subtasks on the terminal device and establishing execution monitoring results; performing execution lag analysis on the subtasks based on the execution monitoring results and establishing execution lag identifiers; establishing optimization constraints based on the execution lag identifiers, and optimizing the execution scheme using the optimization constraints.

[0063] Specifically, this involves monitoring the execution of subtasks on terminal devices. Execution monitoring refers to the real-time tracking and recording of the execution of subtasks by the terminal devices, such as monitoring key indicators like task execution time, computing resource consumption, and data transmission latency. This monitoring data is then organized and stored to create execution monitoring results for subsequent analysis. For example, if a terminal device consistently uses nearly 100% CPU when executing a subtask, while another device executing the same task uses only 50% CPU, this difference is recorded to determine if the task allocation strategy needs adjustment.

[0064] Next, based on the execution monitoring results, execution lag analysis of subtasks is performed, and execution lag indicators are established. The execution efficiency of tasks is then analyzed using the execution monitoring data. Execution lag refers to delays that occur during the execution of subtasks, such as task execution time exceeding expectations or task progress lagging behind the scheduling plan. Execution lag can be caused by various factors, such as insufficient computing resources, network communication latency, or unreasonable task scheduling. Execution lag indicators are used to mark these task delays and optimize task scheduling strategies. For example, if a terminal device experiences task lag due to excessive load, an execution lag indicator can be assigned to that device to remind the task scheduling system to reduce task allocation to that device.

[0065] Then, optimization constraints are established based on the execution lag indicators to optimize the execution plan. Optimization constraints refer to rules set based on the execution lag indicators to optimize task allocation and ensure more efficient task execution. For example, if a device is judged to be overloaded due to execution lag indicators, it can be restricted from receiving new tasks, or some tasks can be reassigned to devices with lower loads. Optimization refers to adjusting the entire task execution plan based on the optimization constraints to improve task execution efficiency. For example, if some devices are found to have persistent execution lag while other devices still have computing power, the optimized plan may reassign tasks, shortening the overall execution time and improving the utilization of computing resources.

[0066] Furthermore, this application also includes: performing a synchronous execution impact analysis of the linked tasks based on the hysteresis identifier, generating hysteresis time constraints based on the impact analysis results; obtaining task requirement constraints of the hysteresis subtasks based on the hysteresis identifier, performing terminal device succession optimization based on the hysteresis time constraints and the task requirement constraints, and optimizing the execution scheme based on the succession optimization results.

[0067] Specifically, the impact analysis of synchronous execution of linked tasks is performed based on lag indicators. Lag indicators are used to identify subtasks whose execution time exceeds expectations. Linked tasks refer to multiple interdependent tasks; for example, the output of one task may be the input of another. Therefore, if one task is delayed, it may affect the execution of subsequent tasks. Synchronous execution impact analysis refers to assessing this impact. For example, if a task is expected to complete within 10 minutes but actually takes 15 minutes, then downstream tasks that depend on that task will also be affected.

[0068] Generating lag time constraints based on the impact analysis results means setting new time constraints. For example, if the execution delay of a certain task affects the overall production process, the time windows of other tasks may be adjusted to reduce the overall delay impact.

[0069] Next, the task requirement constraints of the delayed subtasks are obtained based on the lag identifier. Then, terminal device replacement optimization is performed based on the lag time constraints and task requirement constraints. Delayed subtasks refer to tasks whose execution lag affects the overall progress. Task requirement constraints refer to the computing resources, storage space, network bandwidth, etc., required for these tasks during execution. For example, if a subtask requires high computing power, but the allocated device has low computing power, causing task lag, its computing resource requirement constraints can be adjusted to better match suitable computing resources during the replacement optimization process. Replacement optimization refers to finding terminal devices more suitable for executing these delayed tasks. Possible strategies include transferring tasks to devices with higher computing power, reducing the computational complexity of the tasks, or adjusting the task scheduling order to reduce overall latency. For example, if a device causes task delays due to insufficient computing power, while another device currently has a lower load, the task can be reassigned to the latter to speed up execution.

[0070] Finally, optimization of the execution plan based on the successor optimization result means that after the successor optimization is completed, the overall execution strategy is optimized according to the new task allocation scheme. The goal of optimization is to ensure that all tasks can be completed in the shortest possible time while maximizing the utilization of computing resources. For example, in a smart manufacturing environment, if a task is delayed due to excessive equipment load, successor optimization can find equipment with lower load and reallocate the task, thereby shortening the overall execution time and optimizing production efficiency.

[0071] Furthermore, this application also includes: determining whether the optimization adaptation value of the terminal device that has completed the subtask meets a preset threshold; if the optimization adaptation value of the terminal device that has completed the subtask meets the preset threshold, then the terminal device with the highest optimization adaptation value is output as the successor optimization result.

[0072] Specifically, a terminal device that has completed a subtask refers to a computing node that has finished a particular subtask, completing the current task. There may or may not be subsequent tasks. The process involves determining whether the optimization fit value of the terminal device that has completed the subtask meets a preset threshold. The optimization fit value is an indicator used to measure the quality of task execution by a terminal device, and may include a comprehensive score of multiple factors such as task execution time, computing resource utilization, and energy efficiency. The preset threshold is a set minimum standard used to determine whether the device's performance meets the requirements. For example, if the maximum optimization fit value is 100%, and the preset threshold is 70%, then only devices with an optimization fit value greater than or equal to 70% will be considered qualified execution devices. The process of determining whether the optimization fit value meets the preset threshold helps to filter out better-performing devices and avoid assigning tasks to devices with poor execution capabilities.

[0073] Next, if the optimization adaptation value of the terminal device that has completed the subtask meets the preset threshold, the terminal device with the highest optimization adaptation value will be output as the successor optimization result. If the optimization adaptation value of a terminal device reaches the minimum requirement set by the system, the device with the highest optimization adaptation value among all devices that meet the requirements will be selected as the best candidate to succeed in the task. For example, suppose three terminal devices have completed the subtask and obtained optimization adaptation values ​​of 75%, 85%, and 90% respectively, while the preset threshold is 70%. Then, the terminal device with the highest optimization adaptation value of 90% will be selected as the final successor optimization result. The output of the successor optimization result can be used for subsequent task scheduling. For example, if a new subtask needs to be executed, it can be preferentially assigned to the device with this high adaptation value to ensure more efficient execution.

[0074] Furthermore, this application also includes: if the optimization adaptation value of the terminal device that has completed the subtask cannot meet the preset threshold, then calculate the interruption loss of the terminal device that has not completed the subtask; after compensating the optimization adaptation value of the terminal device that has not completed the subtask based on the interruption loss, reconstruct the successor optimization result.

[0075] Specifically, if the optimization adaptation value of a terminal device that has completed a subtask fails to meet a preset threshold, the interruption loss for the terminal device that failed to complete the subtask is calculated. The optimization adaptation value is an indicator of the quality of task execution by the device. If this value is lower than the preset threshold set by the system, it indicates that the device's performance is unsatisfactory, potentially due to low computational efficiency, excessively long task execution time, or low resource utilization. In this case, the interruption loss for the terminal device that failed to complete the subtask is calculated. A terminal device that failed to complete a subtask refers to a device that failed to complete the task, such as a device whose task was interrupted due to insufficient computing power, network latency, or system failure. Interruption loss is an indicator of the negative impact of a failed task, which may include wasted computing resources, extended task execution time, and cascading effects on other tasks.

[0076] Next, after compensating the terminal devices whose subtasks were not completed based on the interruption loss, the succession optimization results are reconstructed. For devices that did not complete their tasks, the optimization values ​​are adjusted according to their interruption losses to compensate for the impact of task incompleteness. For example, if a device fails to complete a task due to insufficient computing resources, but its communication stability and energy efficiency are high, it may be given some optimization value compensation to more fairly measure its overall capabilities. The compensated optimization values ​​can be used to optimize task scheduling, ensuring that suitable devices are matched more reasonably when tasks are reallocated. Reconstructing the succession optimization results means that after the compensation adjustment, the optimal task succession scheme is recalculated, and the most suitable device is selected to execute the unfinished task.

[0077] In summary, the AIoT cloud-edge-device intelligent control method for industrial IoT provided in this application has the following technical effects: by realizing the technical goal of intelligent task scheduling and optimized allocation under the cloud-edge-device collaborative computing architecture, it achieves the technical effects of improving task execution efficiency, reducing computing latency, optimizing computing resource utilization, and enhancing data security.

[0078] Example 2: Based on the same inventive concept as the AIoT cloud-edge-device intelligent control method for industrial IoT in the foregoing examples, this application also provides an AIoT cloud-edge-device intelligent control system for industrial IoT. Please refer to the appendix. Figure 2 The system includes: a collaborative architecture creation module 11, used to create a cloud-edge-device collaborative architecture, which includes a cloud center, edge computing nodes, and terminal devices, and obtains intelligent manufacturing tasks from the cloud center; an adaptation analysis module 12, used to obtain the device performance and device functions of the terminal devices at the edge computing nodes, perform adaptation analysis on the intelligent manufacturing tasks based on the device performance and device functions, and establish a first adaptation constraint; a task decomposition module 13, used to activate the dynamic task decomposition channel of the cloud center, and use the dynamic task decomposition channel to decompose the intelligent manufacturing tasks and create M sub-tasks; a competitive game analysis module 14, used to use the cloud center to perform task competitive game analysis on the terminal devices for the M sub-tasks, and establish a second adaptation constraint; and a solution generation module 15, used to perform balance analysis on the first adaptation constraint and the second adaptation constraint, and generate an intelligent execution solution for the terminal devices of the intelligent manufacturing tasks.

[0079] Furthermore, the AIoT cloud-edge-device intelligent control system for industrial IoT is also used to: activate the task competition game analysis function in the cloud center, execute task competition game analysis of M sub-tasks, the game characteristics of the competition game analysis function include device adaptability characteristics, historical performance consistency characteristics, communication stability characteristics, and data fidelity guarantee characteristics, and the task competition game analysis function performs game compensation through inter-task collaboration factors.

[0080] Furthermore, the AIoT cloud-edge-device intelligent control system for the Industrial Internet of Things is also used for: uploading the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints of the intelligent manufacturing task; obtaining the highest and lowest computing power of the terminal device, and establishing granular decomposition constraints based on the highest and lowest computing power; and decomposing the intelligent manufacturing task into M sub-tasks based on the granular decomposition constraints, the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints.

[0081] Furthermore, the AIoT cloud-edge-device intelligent control system for industrial IoT is also used for: monitoring the execution of sub-tasks on the terminal device and establishing execution monitoring results; performing execution lag analysis on the sub-tasks based on the execution monitoring results and establishing execution lag identifiers; establishing optimization constraints based on the execution lag identifiers, and optimizing the execution scheme using the optimization constraints.

[0082] Furthermore, the AIoT cloud-edge-device intelligent control system for industrial IoT is also used for: performing synchronous execution impact analysis of linkage tasks based on the hysteresis identifier, generating hysteresis time constraints based on the impact analysis results; obtaining task requirement constraints of hysteresis sub-tasks based on the hysteresis identifier, performing succession optimization of terminal devices based on the hysteresis time constraints and the task requirement constraints, and optimizing the execution scheme based on the succession optimization results.

[0083] Furthermore, the AIoT cloud-edge-device intelligent control system for industrial IoT is also used to: determine whether the optimization adaptation value of the terminal device that has completed the subtask meets the preset threshold; if the optimization adaptation value of the terminal device that has completed the subtask meets the preset threshold, then the terminal device with the highest optimization adaptation value is output as the successor optimization result.

[0084] Furthermore, the AIoT cloud-edge-device intelligent control system for industrial IoT is also used to: if the optimization adaptation value of the terminal device that has completed the subtask cannot meet the preset threshold, calculate the interruption loss of the terminal device that has not completed the subtask; and after compensating the optimization adaptation value of the terminal device that has not completed the subtask according to the interruption loss, reconstruct the successor optimization result.

[0085] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The AIoT cloud-edge-device intelligent control method and specific examples for industrial IoT described in Embodiment 1 above are also applicable to the AIoT cloud-edge-device intelligent control system for industrial IoT in this embodiment. Through the foregoing detailed description of the AIoT cloud-edge-device intelligent control method for industrial IoT, those skilled in the art can clearly understand the AIoT cloud-edge-device intelligent control system for industrial IoT in this embodiment. Therefore, for the sake of brevity, it will not be described in detail here.

[0086] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

[0087] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of this application and its equivalents, this application also intends to include such modifications and variations.

Claims

1. An AIoT cloud-edge-device intelligent control method for the Industrial Internet of Things, characterized in that, include: Create a cloud-edge-device collaborative architecture, which includes a cloud center, edge computing nodes, and terminal devices, and obtain intelligent manufacturing tasks from the cloud center; The device performance and functions of the terminal device are obtained at the edge computing node, and the adaptation analysis of the intelligent manufacturing task is performed based on the device performance and functions to establish the first adaptation constraint. Activate the dynamic task decomposition channel in the cloud center, and use the dynamic task decomposition channel to decompose the intelligent manufacturing task and create M sub-tasks. The cloud center is used to perform task competition game analysis on terminal devices for M sub-tasks, and a second adaptation constraint is established. A balance analysis is performed on the first adaptation constraint and the second adaptation constraint to generate an intelligent execution scheme for the terminal device of the intelligent manufacturing task; The intelligent execution scheme of the terminal device that generates the intelligent manufacturing task includes: The execution of subtasks is monitored on the terminal device, and the execution monitoring results are established. Based on the execution monitoring results, perform execution lag analysis on sub-tasks and establish execution lag identifiers; An optimization constraint is established based on the execution lag indicator, and the execution scheme is optimized using the optimization constraint. The process of using the cloud center to perform task competition game analysis on terminal devices for M sub-tasks includes: Activate the task competition game analysis function in the cloud center, and perform task competition game analysis on M sub-tasks. The game characteristics of the competition game analysis function include device adaptability characteristics, historical performance consistency characteristics, communication stability characteristics, and data fidelity guarantee characteristics. The task competition game analysis function performs game compensation through inter-task collaboration factors.

2. The AIoT cloud-edge-device intelligent control method for industrial IoT as described in claim 1, characterized in that, The method of using the cloud center to perform task competition game analysis on terminal devices for M sub-tasks also includes: Based on the device performance of the terminal device, evaluate the computing power of the terminal device and establish a weak computing device performance protection label. Perform load analysis on the terminal device within a preset period and establish a high load identifier; The equilibrium compensation is performed by using the weak computing device performance protection flag and the high load flag for task competition game analysis, and a second adaptation constraint is established based on the equilibrium compensation result.

3. The AIoT cloud-edge-device intelligent control method for industrial IoT as described in claim 1, characterized in that, The process of decomposing the intelligent manufacturing task using the dynamic task decomposition channel to create M sub-tasks includes: Upload the time constraints, computational complexity constraints, computational requirement constraints, and task association constraints of the intelligent manufacturing task. Obtain the highest and lowest computing power of the terminal device, and establish granularity decomposition constraints based on the highest and lowest computing power. The intelligent manufacturing task is decomposed into M subtasks based on the granularity decomposition constraints, time constraints, computational complexity constraints, computational requirement constraints, and task association constraints.

4. The AIoT cloud-edge-device intelligent control method for industrial IoT as described in claim 1, characterized in that, The step of establishing optimization constraints based on the execution lag identifier and optimizing the execution scheme using the optimization constraints includes: Based on the hysteresis identifier, an impact analysis on the synchronous execution of linked tasks is performed, and hysteresis time constraints are generated based on the impact analysis results. The task requirement constraints of the delayed subtask are obtained based on the hysteresis identifier. The succession optimization of the terminal device is performed based on the hysteresis time constraint and the task requirement constraint. The optimization of the execution scheme is then performed based on the succession optimization result.

5. The AIoT cloud-edge-device intelligent control method for industrial IoT as described in claim 4, characterized in that, The step of optimizing terminal device succession based on the hysteresis time constraint and the task requirement constraint further includes: Determine whether the optimization and adaptation value of the terminal device after the subtask has been completed meets the preset threshold. If the optimization adaptation value of the terminal device that has completed the subtask meets the preset threshold, then the terminal device with the highest optimization adaptation value will be output as the successor optimization result.

6. The AIoT cloud-edge-device intelligent control method for industrial IoT as described in claim 5, characterized in that, The step of determining whether the optimization and adaptation value of the terminal device after the subtask has been completed meets the preset threshold also includes: If the optimization and adaptation value of the terminal device that has completed the subtask cannot meet the preset threshold, then the interruption loss of the terminal device that has not completed the subtask is calculated. After compensating the terminal devices whose subtasks were not completed based on the interruption loss, the successor optimization result is reconstructed.

7. An AIoT cloud-edge-device intelligent control system for the Industrial Internet of Things, characterized in that, The steps for implementing the AIoT cloud-edge-device intelligent control method for industrial IoT as described in any one of claims 1 to 6 include: The collaborative architecture creation module is used to create a cloud-edge-device collaborative architecture, which includes a cloud center, edge computing nodes, and terminal devices, and obtains intelligent manufacturing tasks from the cloud center. The adaptation analysis module is used to obtain the device performance and device functions of the terminal device at the edge computing node, perform adaptation analysis of the smart manufacturing task based on the device performance and device functions, and establish the first adaptation constraint. The task decomposition module is used to activate the dynamic task decomposition channel in the cloud center, and use the dynamic task decomposition channel to decompose the intelligent manufacturing task and create M sub-tasks. The competitive game analysis module is used to perform task competitive game analysis of M sub-tasks on terminal devices using the cloud center, and to establish a second adaptation constraint. The scheme generation module is used to perform a balance analysis on the first adaptation constraint and the second adaptation constraint to generate an intelligent execution scheme for the terminal equipment of the intelligent manufacturing task.

Images