Equipment control method and system for industrial Internet of Things, terminal and storage medium

Abstract

The invention discloses an industrial Internet of Things-oriented equipment control method and system, a terminal and a storage medium, and the method comprises the steps: obtaining industrial equipment information of an industrial equipment layer, obtaining an original equipment operation instruction according to the industrial equipment information, and carrying out the preliminary analysis to obtain a preliminary analysis result; performing structural analysis on the original equipment operation instruction to obtain a structural analysis result, performing semantic analysis to obtain a semantic analysis result, and performing cross-modal similarity calculation to obtain cross-modal similarity; and if the abnormal operation cycle does not exist, performing optimization processing on the original equipment operation instruction to obtain a target equipment operation instruction, and performing equipment control according to the target equipment operation instruction. According to the invention, operation frequency abnormity is captured in real time through a dual-mode cooperative detection architecture, mode offset caused by equipment aging is solved through dynamic optimization of a closed-loop system, stable operation of a high-load scene is ensured by adopting a load feedback mechanism, and the control precision of industrial equipment is improved.

Description

Technical Field

[0001] This invention relates to the field of equipment control technology, and in particular to a device control method, system, terminal, and computer-readable storage medium for the Industrial Internet of Things. Background Technology

[0002] With industrial development, smart factories and automated production lines are increasingly demanding real-time monitoring, fault prediction, and autonomous decision-making. Traditional industrial control systems primarily rely on programmable logic controllers (PLCs) to execute fixed operating procedures and detect anomalies using manually set thresholds. However, facing complex and ever-changing production environments and equipment conditions, existing technologies have the following shortcomings: (1) Hidden resource waste is difficult to detect: The existing structural analysis relies on fixed frequency thresholds, which has a high false alarm rate in the multimodal operation chain of equipment. For example, it may misjudge normal inspection cycles as redundant. Furthermore, text semantic methods cannot process numerical time series data, which leads to the failure of vibration sensing similarity calculation. (2) The contradiction between edge computing power and detection accuracy: full semantic analysis requires high computing power and has high latency on small memory devices, for example, the latency exceeds 5 seconds. Therefore, a single method cannot take into account both real-time performance and accuracy. (3) Industrial protocol and execution flow are separated: the control logic of PLC (Programmable Logic Controller) is separated from the decision-making of intelligent agent, and redundant operation interruption requires manual intervention.

[0003] Therefore, existing technologies still need to be improved and developed. Summary of the Invention

[0004] The main objective of this invention is to provide a device control method, system, terminal, and computer-readable storage medium for the Industrial Internet of Things (IIoT), aiming to solve the problems of high control latency and inaccurate control precision in existing industrial device control methods.

[0005] To achieve the above objectives, the present invention provides a device control method for the Industrial Internet of Things (IIoT), the method comprising the following steps: Obtain industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results; Based on the preliminary analysis results, the original equipment operation instructions are subjected to structural analysis to obtain structural analysis results. Semantic analysis is then performed on the structural analysis results to obtain semantic analysis results. Finally, cross-modal similarity is calculated based on the semantic analysis results to obtain cross-modal similarity. If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, the original device operation command is optimized based on the cross-modal similarity to obtain the target device operation command, and the device is controlled based on the target device operation command.

[0006] Optionally, the device control method for the Industrial Internet of Things (IIoT) includes, in part, acquiring industrial equipment information at the industrial equipment layer, obtaining original device operation commands based on the industrial equipment information, and performing preliminary analysis on the original device operation commands to obtain preliminary analysis results. Specifically, this includes: Obtain industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. Operation instructions are generated based on the industrial equipment operation timestamp and the industrial equipment information to obtain original equipment operation instructions. The original equipment operation instructions are then preliminarily analyzed to obtain preliminary analysis results, which include instruction format verification, operation frequency statistics, and timing continuity statistics.

[0007] Optionally, the device control method for the Industrial Internet of Things (IIoT) further includes, after acquiring industrial equipment information at the industrial equipment layer, obtaining original device operation commands based on the industrial equipment information, and performing preliminary analysis on the original device operation commands to obtain preliminary analysis results: Based on the preliminary analysis results, obtain the equipment operation data of the industrial equipment layer, update the threshold parameters of the equipment operation data, and obtain the threshold optimization parameters. The process of updating the threshold parameters of the device operation data specifically involves: ; in, Optimize parameters for the threshold. For the parameter minimization operator, For threshold parameters, As a loss indicator, For regularization weight parameters, For computational costs on edge devices, This refers to the computing capacity of edge devices.

[0008] Optionally, the device control method for the Industrial Internet of Things, wherein the step of performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results specifically includes: Based on the preliminary analysis results, the corresponding operation subsequences are extracted, and the operation frequency is statistically analyzed based on the subsequences. The operation frequency statistical results are then compared with a preset dynamic threshold to obtain a comparison result. If the comparison result is that the frequency value of the operation frequency statistics result is greater than or equal to the preset dynamic threshold, then the original device operation command is cyclically marked to obtain the marking result; If the comparison result is that the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, then the stack timing sequence is invoked, and the original device operation instructions are lightly screened according to the stack timing sequence to obtain the initial screening result. Based on the threshold optimization parameters, the labeling results, and the initial screening results, the structural analysis results are obtained.

[0009] Optionally, in the aforementioned device control method for the Industrial Internet of Things, the step of performing semantic analysis on the structural analysis results to obtain semantic analysis results specifically includes: The structural analysis results are subjected to data analysis to obtain structured data and unstructured data. The structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. Numerical time series analysis is performed on the structured data to obtain numerical analysis results, and text content analysis is performed on the unstructured data to obtain text analysis results. Semantic analysis results are obtained based on the numerical analysis results and the text analysis results.

[0010] Optionally, the device control method for the Industrial Internet of Things, wherein performing numerical time-series analysis on the structured data to obtain numerical analysis results specifically includes: The structured data is subjected to waveform similarity calculation to obtain waveform similarity calculation results; Numerical analysis results are obtained based on the waveform similarity calculation results; The waveform similarity calculation of the structured data specifically involves: ; in, For sequence and sequence The dynamic time-warped distance between them Align the path The minimum distance in, For sequence The first in One element, For sequence The first in Each element.

[0011] Optionally, in the aforementioned device control method for the Industrial Internet of Things, the step of performing text content analysis on the unstructured data to obtain text analysis results specifically includes: The unstructured data is processed to obtain lightweight data, and vector extraction is performed on the lightweight data to obtain text vectors. The text vectors are subjected to cosine similarity calculation to obtain the cosine similarity calculation result, and the text analysis result is obtained based on the cosine similarity calculation result.

[0012] Optionally, in the aforementioned device control method for the Industrial Internet of Things, the step of calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity specifically includes: The semantic analysis results are used to calculate similarity, and the weights of the multimodal data are configured. The weights of the multimodal data are weighted and fused with the similarity results to obtain a data fusion result. Then, the parameters of the data fusion result are constrained to obtain the cross-modal similarity.

[0013] Optionally, the device control method for the Industrial Internet of Things, wherein the step of performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, performing semantic analysis on the structural analysis results to obtain semantic analysis results, and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity, further includes: An adaptive threshold for triggering a loop is set, and a loop decision is made based on the adaptive threshold and the cross-modal similarity according to the decision function to obtain the loop decision result. Based on the loop decision result, it is determined whether there is an abnormal operation loop in the original device operation instruction. If the loop decision result returns the first value, it is determined that there is an abnormal operation loop in the original device operation instruction; If the loop decision result returns a second value, it is determined that the original device operation instruction does not have an abnormal operation loop.

[0014] Optionally, in the aforementioned device control method for the Industrial Internet of Things, the step of making cyclical decisions based on the adaptive threshold and the cross-modal similarity according to the decision function specifically involves: ; in, Let be the decision function. For cross-modal similarity, This is an adaptive threshold.

[0015] Optionally, the device control method for the Industrial Internet of Things, wherein the step of performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, performing semantic analysis on the structural analysis results to obtain semantic analysis results, and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity, further includes: If the original device operation command is determined to have an abnormal operation loop based on the cross-modal similarity, then the command is generated based on the cross-modal similarity to obtain a loop termination command; The loop termination instruction is encapsulated using a protocol to obtain a target loop termination instruction, and the abnormal operation loop is terminated according to the target loop termination instruction to obtain a loop termination result.

[0016] Optionally, the device control method for the Industrial Internet of Things, wherein if it is determined based on the cross-modal similarity that the original device operation command does not have an abnormal operation loop, then the original device operation command is optimized based on the cross-modal similarity to obtain a target device operation command, and device control is performed based on the target device operation command, specifically includes: If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, then a dynamic scheduling strategy is generated based on the cross-modal similarity. Resource reallocation is performed according to the dynamic scheduling strategy to obtain resource allocation results. The original equipment operation instructions are then optimized based on the resource allocation results to obtain target equipment operation instructions. Equipment control is then performed based on the target equipment operation instructions.

[0017] Optionally, in the aforementioned device control method for the Industrial Internet of Things, the step of performing resource reallocation processing according to the dynamic scheduling strategy specifically includes: ; in, To allocate new resources, For the recycling of idle resources, This is the allocation ratio coefficient. This represents the task priority gradient.

[0018] Furthermore, to achieve the above objectives, the present invention also provides a device control system for the Industrial Internet of Things (IIoT), wherein the device control system for the Industrial Internet of Things includes: The operation instruction generation module is used to acquire industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results. The operation instruction analysis module is used to perform structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, perform semantic analysis on the structural analysis results to obtain semantic analysis results, and perform cross-modal similarity calculation based on the semantic analysis results to obtain cross-modal similarity. The operation instruction optimization module is used to optimize the original device operation instruction according to the cross-modal similarity if it is determined that there is no abnormal operation loop in the original device operation instruction, to obtain the target device operation instruction, and to control the device according to the target device operation instruction.

[0019] Optionally, in the aforementioned equipment control system for the Industrial Internet of Things, the operation instruction generation module includes: The data acquisition unit is used to acquire industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. The data analysis unit is used to generate operation instructions based on the operation timestamp of the industrial equipment and the information of the industrial equipment, obtain the original equipment operation instructions, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, wherein the preliminary analysis results include instruction format verification, operation frequency statistics and timing continuity statistics.

[0020] Optionally, in the aforementioned equipment control system for the Industrial Internet of Things, the operation instruction analysis module includes: The data statistics unit is used to extract the corresponding operation sub-sequences based on the preliminary analysis results, perform operation frequency statistics based on the sub-sequences, obtain the operation frequency statistics results, and compare the operation frequency statistics results with a preset dynamic threshold to obtain a comparison result. A cyclic marking unit is used to cyclically mark the original device operation command if the comparison result is greater than or equal to the frequency value of the operation frequency statistics result, thereby obtaining a marking result; The data screening unit is used to call the stack timing sequence if the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, and perform a lightweight screening of the original device operation instructions according to the stack timing sequence to obtain the screening result. The structural analysis unit is used to obtain structural analysis results based on the threshold optimization parameters, the labeling results, and the initial screening results; The data analysis unit is used to perform data analysis on the structural analysis results to obtain structured data and unstructured data, wherein the structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. The text analysis unit is used to perform numerical time series analysis on the structured data to obtain numerical analysis results, and to perform text content analysis on the unstructured data to obtain text analysis results, and to obtain semantic analysis results based on the numerical analysis results and the text analysis results. A similarity calculation unit is used to calculate the similarity of the semantic analysis results, obtain the similarity results, and configure the multimodal data weights; The data fusion unit is used to perform weighted fusion of the multimodal data weights and the similarity results to obtain a data fusion result, and to impose parameter constraints on the data fusion result to obtain cross-modal similarity.

[0021] Optionally, in the aforementioned equipment control system for the Industrial Internet of Things, the operation instruction optimization module includes: The scheduling strategy generation unit is used to generate a dynamic scheduling strategy based on the cross-modal similarity if it is determined that the original device operation instruction does not have an abnormal operation loop based on the cross-modal similarity. The resource reallocation unit is used to perform resource reallocation processing according to the dynamic scheduling strategy to obtain resource allocation results, optimize the original equipment operation instructions according to the resource allocation results to obtain target equipment operation instructions, and perform equipment control according to the target equipment operation instructions.

[0022] Optionally, the device control system for the Industrial Internet of Things (IIoT) further includes a cyclic decision-making module, which includes: The loop judgment unit is used to set an adaptive threshold for triggering the loop, perform loop decision on the adaptive threshold and the cross-modal similarity according to the decision function, obtain the loop decision result, and determine whether there is an abnormal operation loop in the original device operation instruction based on the loop decision result; The first loop determination unit is used to determine that there is an abnormal operation loop in the original device operation instruction if the loop decision result returns a first value. The second loop determination unit is used to determine that the original device operation instruction does not have an abnormal operation loop if the loop decision result returns a second value.

[0023] Furthermore, to achieve the above objectives, the present invention also provides a terminal, wherein the terminal includes: a memory, a processor, and a device control program for the Industrial Internet of Things (IIoT) stored in the memory and executable on the processor, wherein when the device control program for the IIoT is executed by the processor, it implements the steps of the device control method for the IIoT as described above.

[0024] Furthermore, to achieve the above objectives, the present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a device control program for the Industrial Internet of Things (IIoT), and when the IIoT-oriented device control program is executed by a processor, it implements the steps of the device control method for the IIoT as described above.

[0025] In this invention, industrial equipment information at the industrial equipment layer is acquired, and original equipment operation instructions are obtained based on the industrial equipment information. These original equipment operation instructions are then preliminarily analyzed to obtain preliminary analysis results. Based on the preliminary analysis results, structural analysis is performed on the original equipment operation instructions to obtain structural analysis results. Semantic analysis is then performed on the structural analysis results to obtain semantic analysis results. Cross-modal similarity is calculated based on the semantic analysis results to obtain cross-modal similarity. If the cross-modal similarity indicates that the original equipment operation instructions do not have abnormal operation loops, then the original equipment operation instructions are optimized based on the cross-modal similarity to obtain target equipment operation instructions. Equipment control is then performed based on the target equipment operation instructions. This invention uses a dual-mode collaborative detection architecture to capture abnormal operation frequencies in real time. It also addresses mode shifts caused by equipment aging through dynamic optimization of the closed-loop system and employs a load feedback mechanism to ensure stable operation in high-load scenarios, thereby improving the control accuracy of industrial equipment. Attached Figure Description

[0026] Figure 1 This is a flowchart of a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention; Figure 2 This is a flowchart illustrating the specific implementation process of step S10 in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention. Figure 3 This is a flowchart illustrating the specific implementation process of step S20 in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention. Figure 4 This is a flowchart of numerical time series analysis of structured data in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention; Figure 5 This is a flowchart of text content analysis of unstructured data in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention; Figure 6 This is a flowchart illustrating the specific implementation process of step S30 in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention. Figure 7 This is a flowchart illustrating the specific implementation process of step S40 in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention. Figure 8 This is a flowchart illustrating the specific implementation process of step S50 in a preferred embodiment of the device control method for the Industrial Internet of Things of the present invention. Figure 9 This is a schematic diagram of a device control system for the Industrial Internet of Things according to the present invention; Figure 10 This is another structural schematic diagram of the device control system of the present invention for the Industrial Internet of Things; Figure 11 This is a schematic diagram of the data analysis module in the equipment control system of the present invention, which is oriented towards the Industrial Internet of Things; Figure 12 This is a schematic diagram of the model training module in the device control system for the Industrial Internet of Things according to the present invention; Figure 13 This is a schematic diagram of the image super-resolution module in the device control system of the present invention for the Industrial Internet of Things; Figure 14 This is a schematic diagram of the cyclic decision-making module in the equipment control system of the present invention for the Industrial Internet of Things; Figure 15 This is a structural diagram of a preferred embodiment of the terminal of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0028] Traditional industrial control systems primarily rely on programmable logic controllers (PLCs) to execute fixed operational procedures and use manually set thresholds for anomaly detection. However, given the complex and ever-changing production environments and equipment conditions, existing technologies have the following shortcomings: (1) Hidden resource waste is difficult to detect: The existing structural analysis relies on fixed frequency thresholds, which has a high false alarm rate in the multimodal operation chain of equipment. For example, it may misjudge normal inspection cycles as redundant. Furthermore, text semantic methods cannot process numerical time series data, which leads to the failure of vibration sensing similarity calculation. (2) The contradiction between edge computing power and detection accuracy: full semantic analysis requires high computing power and has high latency on small memory devices, for example, the latency exceeds 5 seconds. Therefore, a single method cannot take into account both real-time performance and accuracy. (3) Industrial protocol and execution flow are separated: the control logic of PLC (Programmable Logic Controller) is separated from the decision-making of intelligent agent, and redundant operation interruption requires manual intervention.

[0029] Therefore, this invention proposes a device control method for the Industrial Internet of Things (IIoT) to overcome the problems of high control delay and inaccurate control precision in existing industrial device control methods.

[0030] The preferred embodiment of the device control method for the Industrial Internet of Things (IIoT) of the present invention, such as... Figure 1 As shown, the device control method for the Industrial Internet of Things includes the following steps: Step S10: Obtain industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results.

[0031] Specifically, in this embodiment of the invention, multiple low-resolution face image streams from a HarmonyOS device are acquired, and dynamic resolution input is supported (e.g., 16×16 to 256×256 pixels). After acquiring multiple low-resolution face image streams, dynamic quality assessment and filtering processing is required for all low-resolution face image streams. This includes face sharpness filtering, face pose angle filtering, and face occlusion ratio filtering, retaining only images with quality scores higher than the dynamic threshold to ensure the quality of the training set. Then, a normalization layer preprocessing is used to ensure the stability of subsequent calculations.

[0032] like Figure 2 As shown, step S10 specifically includes: Step S11: Obtain industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log.

[0033] Specifically, step S11 is implemented as follows: The process involves acquiring industrial equipment information and operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. Operation instructions are generated based on the industrial equipment operation timestamps and the industrial equipment information to obtain original equipment operation instructions. Preliminary analysis of these original equipment operation instructions yields preliminary analysis results, including instruction format verification, operation frequency statistics, and timing continuity statistics.

[0034] Step S12: Generate operation instructions based on the industrial equipment operation timestamp and the industrial equipment information to obtain original equipment operation instructions, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, wherein the preliminary analysis results include instruction format verification, operation frequency statistics and timing continuity statistics.

[0035] Specifically, step S12 is implemented as follows: Operation instructions are generated based on the industrial equipment operation timestamp and the industrial equipment information to obtain original equipment operation instructions. The original equipment operation instructions are then preliminarily analyzed to obtain preliminary analysis results, which include instruction format verification, operation frequency statistics, and time sequence continuity statistics. By identifying the uniqueness of tasks and maintaining the operation hierarchy, basic topology data is provided for subsequent structural analysis.

[0036] Furthermore, after obtaining the preliminary analysis results, the equipment operation data of the industrial equipment layer is obtained based on the preliminary analysis results. The threshold parameters of the equipment operation data are then updated to obtain the threshold optimization parameters, the corresponding expression of which is: ; in, Optimize parameters for the threshold. For the parameter minimization operator, For threshold parameters, As a loss indicator, For regularization weight parameters, For computational costs on edge devices, This refers to the computing capacity of edge devices.

[0037] Step S20: Perform structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results; perform semantic analysis on the structural analysis results to obtain semantic analysis results; and calculate cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity.

[0038] Specifically, after obtaining the preliminary analysis results, it is necessary to perform structural analysis on the original device operation instructions based on the preliminary analysis results. Specifically, this involves extracting corresponding operation sub-sequences based on the preliminary analysis results, performing operation frequency statistics on the sub-sequences, and comparing the operation frequency statistics results with a preset dynamic threshold to obtain a comparison result. If the comparison result indicates that the frequency value of the operation frequency statistics is greater than or equal to the preset dynamic threshold, then the original device operation instructions are cyclically marked to obtain a marking result. If the comparison result indicates that the frequency value of the operation frequency statistics is less than the preset dynamic threshold, then a stack timing sequence is invoked, and the original device operation instructions are lightly screened based on the stack timing sequence to obtain a preliminary screening result. Finally, based on the threshold optimization parameters, the marking result, and the preliminary screening result, the structural analysis result is obtained.

[0039] Subsequently, data analysis is performed on the structural analysis results to obtain structured and unstructured data. The structured data includes vibration signal data and equipment timing data, while the unstructured data includes log text data and control command data. Numerical time-series analysis is performed on the structured data to obtain numerical analysis results, and text content analysis is performed on the unstructured data to obtain text analysis results. Semantic analysis results are then obtained based on the numerical analysis results and the text analysis results.

[0040] Finally, similarity is calculated on the semantic analysis results to obtain similarity results, and multimodal data weights are configured; the multimodal data weights and the similarity results are weighted and fused to obtain data fusion results, and parameter constraints are applied to the data fusion results to obtain cross-modal similarity.

[0041] like Figure 3 As shown, step S20 specifically includes: Step S21: Extract the corresponding operation subsequences based on the preliminary analysis results, perform operation frequency statistics based on the subsequences, obtain the operation frequency statistics results, and compare the operation frequency statistics results with a preset dynamic threshold to obtain the comparison results.

[0042] Specifically, step S21 is implemented as follows: Based on the preliminary analysis results, the corresponding operation subsequences are extracted, and the operation frequency is statistically analyzed based on the subsequences. The operation frequency statistical results are then compared with a preset dynamic threshold to obtain a comparison result. This invention detects repetitive operation patterns through a real-time sliding window. When the frequency of a subsequence exceeds the preset dynamic threshold, it is marked as a suspected cycle.

[0043] Step S22: If the comparison result is that the frequency value of the operation frequency statistics result is greater than or equal to the preset dynamic threshold, then the original device operation command is cyclically marked to obtain the marking result.

[0044] Specifically, step S22 is implemented as follows: If the comparison result is that the frequency value of the operation frequency statistics is greater than or equal to the preset dynamic threshold, it is marked as a suspected loop, that is, the original device operation command is marked as a loop to obtain the marking result.

[0045] Step S23: If the comparison result is that the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, then the stack timing sequence is invoked, and the original device operation instructions are lightly screened according to the stack timing sequence to obtain the initial screening result.

[0046] Specifically, step S23 is implemented as follows: If the comparison result is that the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, then the stack timing sequence is invoked, and the original device operation instructions are lightly screened according to the stack timing sequence to obtain the initial screening result.

[0047] Step S24: Obtain the structural analysis results based on the threshold optimization parameters, the labeling results, and the initial screening results.

[0048] Specifically, step S24 is implemented as follows: Based on the threshold optimization parameters, the labeling results, and the initial screening results, the structural analysis results are obtained.

[0049] Step S25: Perform data analysis on the structural analysis results to obtain structured data and unstructured data. The structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data.

[0050] Specifically, step S25 is implemented as follows: The structural analysis results are then subjected to data analysis to obtain structured and unstructured data. The structured data includes vibration signal data and equipment timing data, while the unstructured data includes log text data and control command data. The purpose of the data analysis is to perform different processing on different types of data. Numerical timing analysis is required for structured data, while text content analysis is required for unstructured data.

[0051] Step S26: Perform numerical time series analysis on the structured data to obtain numerical analysis results, perform text content analysis on the unstructured data to obtain text analysis results, and obtain semantic analysis results based on the numerical analysis results and the text analysis results.

[0052] Specifically, step S26 is implemented as follows: The process of performing numerical time series analysis on the structured data to obtain the numerical analysis results is as follows: Figure 4 As shown, specifically, waveform similarity calculation is performed on the structured data to obtain waveform similarity calculation results; Numerical analysis results are obtained based on the waveform similarity calculation results; The waveform similarity calculation of the structured data specifically involves: ; in, For sequence and sequence The dynamic time-warped distance between them Align the path The minimum distance in, For sequence The first in One element, For sequence The first in Each element.

[0053] The process of performing numerical time series analysis on the structured data to obtain the numerical analysis results is as follows: Figure 5 As shown, specifically, the unstructured data is subjected to lightweight processing to obtain lightweight data, and the lightweight data is subjected to vector extraction to obtain text vectors; The text vectors are subjected to cosine similarity calculation to obtain the cosine similarity calculation result, and the text analysis result is obtained based on the cosine similarity calculation result.

[0054] Step S27: Calculate the similarity of the semantic analysis results to obtain the similarity results, and configure the multimodal data weights, wherein the multimodal data weights can be dynamically configured by the cloud node based on the device type.

[0055] Specifically, step S27 is implemented as follows: The semantic analysis results are used to calculate similarity, and the weights of the multimodal data are configured.

[0056] Step S28: Weight the multimodal data weights and the similarity results to obtain a data fusion result, and impose parameter constraints on the data fusion result to obtain cross-modal similarity.

[0057] Specifically, step S28 is implemented as follows: The weights of the multimodal data are weighted and fused with the similarity results to obtain a data fusion result. Then, the parameters of the data fusion result are constrained to obtain the cross-modal similarity.

[0058] Step S30: If it is determined that the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, then the original device operation command is optimized based on the cross-modal similarity to obtain the target device operation command, and the device is controlled based on the target device operation command.

[0059] Specifically, if it is determined that the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, a dynamic scheduling strategy is generated based on the cross-modal similarity; resource reallocation is performed based on the dynamic scheduling strategy to obtain resource allocation results, and the original device operation command is optimized based on the resource allocation results to obtain target device operation commands, and device control is performed based on the target device operation commands.

[0060] like Figure 6 As shown, step S30 specifically includes: Step S31: If it is determined that the original device operation instruction does not have an abnormal operation loop based on the cross-modal similarity, then a dynamic scheduling strategy is generated based on the cross-modal similarity.

[0061] Specifically, step S31 is implemented as follows: If the original device operation instructions are determined to have no abnormal operation loop based on the cross-modal similarity, then resource reallocation is required, i.e., a dynamic scheduling strategy is generated based on the cross-modal similarity.

[0062] Step S32: Perform resource reallocation processing according to the dynamic scheduling strategy to obtain resource allocation results, optimize the original device operation instructions according to the resource allocation results to obtain target device operation instructions, and perform device control according to the target device operation instructions.

[0063] Specifically, step S32 is implemented as follows: After generating the dynamic scheduling policy, resource reallocation is performed according to the dynamic scheduling policy to obtain the resource allocation result, the corresponding expression of which is: ; in, To allocate new resources, For the recycling of idle resources, This is the allocation ratio coefficient. This represents the task priority gradient. Then, based on the resource allocation results, the original device operation instructions are optimized to obtain the target device operation instructions, and device control is performed according to these target device operation instructions.

[0064] Furthermore, such as Figure 7 As shown, based on the preliminary analysis results, structural analysis is performed on the original device operation instructions to obtain structural analysis results. Semantic analysis is then performed on the structural analysis results to obtain semantic analysis results. Cross-modal similarity is calculated based on the semantic analysis results to obtain cross-modal similarity. The process further includes: Step S41: Set an adaptive threshold for triggering a loop, perform a loop decision on the adaptive threshold and the cross-modal similarity according to the decision function, obtain the loop decision result, and determine whether there is an abnormal operation loop in the original device operation instruction based on the loop decision result; Step S42: If the loop decision result returns the first value, it is determined that the original device operation instruction has an abnormal operation loop; Step S43: If the loop decision result returns a second value, it is determined that the original device operation instruction does not have an abnormal operation loop.

[0065] Specifically, after calculating the cross-modal similarity, it is necessary to determine whether there is an abnormal operation loop in the original device operation instructions. In this invention, loop confirmation is performed through loop confirmation decision. Specifically, an adaptive threshold for triggering the loop is set, and a loop decision is made based on the adaptive threshold and the cross-modal similarity according to the decision function to obtain the loop decision result. The corresponding expression is: ; in, Let be the decision function. For cross-modal similarity, This is an adaptive threshold.

[0066] Subsequently, based on the cyclic decision result, it is determined whether the original device operation instruction has an abnormal operation cycle; if the cyclic decision result returns a first value (e.g., 1), it is determined that the original device operation instruction has an abnormal operation cycle; if the cyclic decision result returns a second value (e.g., 0), it is determined that the original device operation instruction does not have an abnormal operation cycle; the present invention also introduces a load feedback mechanism, which automatically reduces the adaptive threshold when the edge device is under high load, so as to accelerate the determination and thus balance accuracy and real-time performance.

[0067] Furthermore, such as Figure 8As shown, the process involves performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, performing semantic analysis on the structural analysis results to obtain semantic analysis results, and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity. The process further includes: Step S51: If the original device operation command is determined to have an abnormal operation loop based on the cross-modal similarity, then the command is generated based on the cross-modal similarity to obtain a loop termination command; Step S52: Encapsulate the loop termination instruction using a protocol to obtain a target loop termination instruction, and terminate the abnormal operation loop according to the target loop termination instruction to obtain a loop termination result.

[0068] Specifically, if the original device operation command is determined to have an abnormal operation loop based on the cross-modal similarity, then the command needs to be terminated by a blocking command. Specifically, the command is generated based on the cross-modal similarity to obtain a loop termination command; the loop termination command is then encapsulated using a protocol to obtain a target loop termination command, the corresponding expression of which is: ; in, This is a loop termination instruction. A unique identifier for the device. For opcodes, This is a hexadecimal opcode, indicating a termination instruction. The execution context ID is then used. Following this, the abnormal operation loop is terminated according to the target loop termination instruction, resulting in a loop termination result.

[0069] The technical effects that this invention can bring are as follows: (1) Structural layer protection and semantic layer precision judgment are carried out through the modular collaborative detection architecture. Structural layer protection is achieved by capturing abnormal operation frequency in real time through a sliding window and filtering 95% of normal data flow through the edge layer. Semantic layer precision judgment is achieved by activating cross-modal analysis only for suspicious data, which reduces computing power consumption. (2) By blocking the direct generation of OPC UA commands through the protocol deep embedding mechanism, it can seamlessly connect with multiple third-party industrial control systems. The resource recycling strategy used can instantly improve the computing power of high-priority tasks through gradient allocation algorithm. (3) By dynamically optimizing the closed-loop system, the cloud node updates the parameter package every 24 hours to solve the mode shift caused by equipment aging. The load feedback mechanism also ensures stable operation in high-load scenarios.

[0070] Furthermore, such as Figure 9As shown, based on the above-described device control method for the Industrial Internet of Things (IIoT), the present invention also provides a device control system for the Industrial Internet of Things, wherein the device control system for the Industrial Internet of Things includes: The operation instruction generation module 50 is used to acquire industrial equipment information of the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results. The operation instruction analysis module 60 is used to perform structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, perform semantic analysis on the structural analysis results to obtain semantic analysis results, and perform cross-modal similarity calculation based on the semantic analysis results to obtain cross-modal similarity. The operation instruction optimization module 70 is used to optimize the original device operation instruction according to the cross-modal similarity if it is determined that there is no abnormal operation loop in the original device operation instruction, to obtain the target device operation instruction, and to control the device according to the target device operation instruction.

[0071] like Figures 10 to 13 As shown in this embodiment of the invention, another embodiment of the device control system for the Industrial Internet of Things (IIoT) includes, in this embodiment, the operation instruction generation module 50 comprising: The data acquisition unit 501 is used to acquire industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. The data analysis unit 502 is used to generate operation instructions based on the operation timestamp of the industrial equipment and the information of the industrial equipment, obtain the original equipment operation instructions, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, wherein the preliminary analysis results include instruction format verification, operation frequency statistics and timing continuity statistics.

[0072] In this embodiment, the operation instruction analysis module 60 includes: The data statistics unit 601 is used to extract the corresponding operation subsequences based on the preliminary analysis results, perform operation frequency statistics based on the subsequences, obtain the operation frequency statistics results, and compare the operation frequency statistics results with a preset dynamic threshold to obtain a comparison result. The cyclic marking unit 602 is used to cyclically mark the original device operation command if the comparison result is that the frequency value of the operation frequency statistics result is greater than or equal to the preset dynamic threshold, so as to obtain the marking result; The data screening unit 603 is used to call the stack timing sequence if the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, and perform a lightweight screening of the original device operation instructions according to the stack timing sequence to obtain the screening result. The structural analysis unit 604 is used to obtain structural analysis results based on the threshold optimization parameters, the labeling results, and the initial screening results; The data analysis unit 605 is used to perform data analysis on the structural analysis results to obtain structured data and unstructured data, wherein the structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. The text analysis unit 606 is used to perform numerical time series analysis on the structured data to obtain numerical analysis results, and to perform text content analysis on the unstructured data to obtain text analysis results, and to obtain semantic analysis results based on the numerical analysis results and the text analysis results. The similarity calculation unit 607 is used to calculate the similarity of the semantic analysis results, obtain the similarity results, and configure the multimodal data weights. The data fusion unit 608 is used to perform weighted fusion of the multimodal data weights and the similarity results to obtain a data fusion result, and to impose parameter constraints on the data fusion result to obtain cross-modal similarity.

[0073] In this embodiment, the operation instruction optimization module 70 includes: The scheduling strategy generation unit 701 is used to generate a dynamic scheduling strategy based on the cross-modal similarity if it is determined that the original device operation instruction does not have an abnormal operation loop based on the cross-modal similarity. The resource reallocation unit 702 is used to perform resource reallocation processing according to the dynamic scheduling strategy to obtain resource allocation results, optimize the original equipment operation instructions according to the resource allocation results to obtain target equipment operation instructions, and perform equipment control according to the target equipment operation instructions.

[0074] like Figure 14 As shown, in this embodiment, the device control system for the Industrial Internet of Things further includes a cyclic decision module 80, which includes: The loop judgment unit 801 is used to set an adaptive threshold for triggering the loop, perform loop decision on the adaptive threshold and the cross-modal similarity according to the decision function, obtain the loop decision result, and determine whether there is an abnormal operation loop in the original device operation instruction based on the loop decision result; The first loop determination unit 802 is used to determine that there is an abnormal operation loop in the original device operation instruction if the loop decision result returns a first value. The second loop determination unit 803 is used to determine that there is no abnormal operation loop in the original device operation instruction if the loop decision result returns a second value.

[0075] This invention captures abnormal operation frequency in real time through a dual-mode collaborative detection architecture, solves mode shift caused by equipment aging by dynamically optimizing the closed-loop system, and uses a load feedback mechanism to ensure stable operation in high-load scenarios, thereby improving the control accuracy of industrial equipment.

[0076] Furthermore, such as Figure 15 As shown, based on the above-mentioned device control method and system for the Industrial Internet of Things, the present invention also provides a terminal, which includes a processor 10, a memory 20 and a display 30. Figure 15 Only some of the terminal components are shown; however, it should be understood that it is not required to implement all of the components shown, and more or fewer components may be implemented instead.

[0077] In some embodiments, the memory 20 may be an internal storage unit of the terminal, such as a hard drive or memory. In other embodiments, the memory 20 may be an external storage device of the terminal, such as a plug-in hard drive, smart media card (SMC), secure digital card (SD), flash card, etc. Further, the memory 20 may include both internal and external storage devices. The memory 20 is used to store application software and various types of data installed on the terminal, such as the program code installed on the terminal. The memory 20 can also be used to temporarily store data that has been output or will be output. In one embodiment, the memory 20 stores a device control program 40 for the Industrial Internet of Things (IIoT), which can be executed by the processor 10 to implement the device control method for the IIoT described in this application.

[0078] In some embodiments, the processor 10 may be a central processing unit (CPU), a microprocessor, or other data processing chip, used to run program code stored in the memory 20 or process data, such as executing the device control method for the Industrial Internet of Things.

[0079] In some embodiments, the display 30 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touchscreen. The display 30 is used to display information on the terminal and to display a visual user interface. The terminal's processor 10, memory 20, and display 30 communicate with each other via a system bus.

[0080] In one embodiment, when the processor 10 executes the device control program 40 for the Industrial Internet of Things stored in the memory 20, the following steps are performed: Obtain industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results; Based on the preliminary analysis results, the original equipment operation instructions are subjected to structural analysis to obtain structural analysis results. Semantic analysis is then performed on the structural analysis results to obtain semantic analysis results. Finally, cross-modal similarity is calculated based on the semantic analysis results to obtain cross-modal similarity. If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, the original device operation command is optimized based on the cross-modal similarity to obtain the target device operation command, and the device is controlled based on the target device operation command.

[0081] Specifically, the process of acquiring industrial equipment information from the industrial equipment layer, obtaining original equipment operation instructions based on the industrial equipment information, and performing preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results includes: Obtain industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. Operation instructions are generated based on the industrial equipment operation timestamp and the industrial equipment information to obtain original equipment operation instructions. The original equipment operation instructions are then preliminarily analyzed to obtain preliminary analysis results, which include instruction format verification, operation frequency statistics, and timing continuity statistics.

[0082] The process includes acquiring industrial equipment information from the industrial equipment layer, obtaining original equipment operation instructions based on the industrial equipment information, performing preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, and then further including: Based on the preliminary analysis results, obtain the equipment operation data of the industrial equipment layer, update the threshold parameters of the equipment operation data, and obtain the threshold optimization parameters. The process of updating the threshold parameters of the device operation data specifically involves: ; in, Optimize parameters for the threshold. For the parameter minimization operator, For threshold parameters, As a loss indicator, For regularization weight parameters, For computational costs on edge devices, This refers to the computing capacity of edge devices.

[0083] Specifically, the step of performing structural analysis on the original equipment operation instructions based on the preliminary analysis results to obtain structural analysis results includes: Based on the preliminary analysis results, the corresponding operation subsequences are extracted, and the operation frequency is statistically analyzed based on the subsequences. The operation frequency statistical results are then compared with a preset dynamic threshold to obtain a comparison result. If the comparison result is that the frequency value of the operation frequency statistics result is greater than or equal to the preset dynamic threshold, then the original device operation command is cyclically marked to obtain the marking result; If the comparison result is that the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, then the stack timing sequence is invoked, and the original device operation instructions are lightly screened according to the stack timing sequence to obtain the initial screening result. Based on the threshold optimization parameters, the labeling results, and the initial screening results, the structural analysis results are obtained.

[0084] Specifically, performing semantic analysis on the structural analysis results to obtain semantic analysis results involves: The structural analysis results are subjected to data analysis to obtain structured data and unstructured data. The structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. Numerical time series analysis is performed on the structured data to obtain numerical analysis results, and text content analysis is performed on the unstructured data to obtain text analysis results. Semantic analysis results are obtained based on the numerical analysis results and the text analysis results.

[0085] Specifically, the step of performing numerical time-series analysis on the structured data to obtain numerical analysis results includes: The structured data is subjected to waveform similarity calculation to obtain waveform similarity calculation results; Numerical analysis results are obtained based on the waveform similarity calculation results; The waveform similarity calculation of the structured data specifically involves: ; in, For sequence and sequence The dynamic time-warped distance between them Align the path The minimum distance in, For sequence The first in One element, For sequence The first in Each element.

[0086] Specifically, the text content analysis of the unstructured data to obtain the text analysis results includes: The unstructured data is processed to obtain lightweight data, and vector extraction is performed on the lightweight data to obtain text vectors. The text vectors are subjected to cosine similarity calculation to obtain the cosine similarity calculation result, and the text analysis result is obtained based on the cosine similarity calculation result.

[0087] Specifically, the calculation of cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity includes: The semantic analysis results are used to calculate similarity, and the weights of the multimodal data are configured. The weights of the multimodal data are weighted and fused with the similarity results to obtain a data fusion result. Then, the parameters of the data fusion result are constrained to obtain the cross-modal similarity.

[0088] The process includes: performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results; performing semantic analysis on the structural analysis results to obtain semantic analysis results; and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity. The process further includes: An adaptive threshold for triggering a loop is set, and a loop decision is made based on the adaptive threshold and the cross-modal similarity according to the decision function to obtain the loop decision result. Based on the loop decision result, it is determined whether there is an abnormal operation loop in the original device operation instruction. If the loop decision result returns the first value, it is determined that there is an abnormal operation loop in the original device operation instruction; If the loop decision result returns a second value, it is determined that the original device operation instruction does not have an abnormal operation loop.

[0089] Specifically, the step of making iterative decisions on the adaptive threshold and the cross-modal similarity based on the decision function involves: ; in, Let be the decision function. For cross-modal similarity, This is an adaptive threshold.

[0090] The process includes: performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results; performing semantic analysis on the structural analysis results to obtain semantic analysis results; and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity. The process further includes: If the original device operation command is determined to have an abnormal operation loop based on the cross-modal similarity, then the command is generated based on the cross-modal similarity to obtain a loop termination command; The loop termination instruction is encapsulated using a protocol to obtain a target loop termination instruction, and the abnormal operation loop is terminated according to the target loop termination instruction to obtain a loop termination result.

[0091] Specifically, if the cross-modal similarity determines that the original device operation command does not have an abnormal operation loop, then the original device operation command is optimized based on the cross-modal similarity to obtain a target device operation command, and device control is performed based on the target device operation command. If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, then a dynamic scheduling strategy is generated based on the cross-modal similarity. Resource reallocation is performed according to the dynamic scheduling strategy to obtain resource allocation results. The original equipment operation instructions are then optimized based on the resource allocation results to obtain target equipment operation instructions. Equipment control is then performed based on the target equipment operation instructions.

[0092] Specifically, the resource reallocation process based on the dynamic scheduling strategy includes: ; in, To allocate new resources, For the recycling of idle resources, This is the allocation ratio coefficient. This represents the task priority gradient.

[0093] The present invention also provides a computer-readable storage medium, wherein the computer-readable storage medium stores a device control program for the Industrial Internet of Things (IIoT), and the IIoT-oriented device control program, when executed by a processor, implements the steps of the device control method for the Industrial Internet of Things as described above.

[0094] In summary, this invention provides a device control method, system, terminal, and computer-readable storage medium for the Industrial Internet of Things (IIoT). The method includes acquiring industrial equipment information at the industrial equipment layer; obtaining original device operation instructions based on the industrial equipment information; performing preliminary analysis on the original device operation instructions to obtain preliminary analysis results; performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results; performing semantic analysis on the structural analysis results to obtain semantic analysis results; calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity; if it is determined based on the cross-modal similarity that the original device operation instructions do not have abnormal operation loops, then optimizing the original device operation instructions based on the cross-modal similarity to obtain target device operation instructions, and controlling the device based on the target device operation instructions. This invention uses a dual-mode collaborative detection architecture to capture abnormal operation frequencies in real time, dynamically optimizes the closed-loop system to solve mode shifts caused by equipment aging, and employs a load feedback mechanism to ensure stable operation in high-load scenarios, thereby improving the control accuracy of industrial equipment.

[0095] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or terminal that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or terminal. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or terminal that includes that element.

[0096] Of course, those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware (such as a processor, controller, etc.). The program can be stored in a computer-readable storage medium, and when executed, it can include the processes described in the above method embodiments. The computer-readable storage medium can be a memory, magnetic disk, optical disk, etc.

[0097] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. A device control method for the Industrial Internet of Things (IIoT), characterized in that, The device control method for the Industrial Internet of Things includes: Obtain industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results; Based on the preliminary analysis results, the original equipment operation instructions are subjected to structural analysis to obtain structural analysis results. Semantic analysis is then performed on the structural analysis results to obtain semantic analysis results. Finally, cross-modal similarity is calculated based on the semantic analysis results to obtain cross-modal similarity. If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, the original device operation command is optimized based on the cross-modal similarity to obtain the target device operation command, and the device is controlled based on the target device operation command.

2. The device control method for the Industrial Internet of Things according to claim 1, characterized in that, The process of acquiring industrial equipment information at the industrial equipment layer, obtaining original equipment operation instructions based on the industrial equipment information, and performing preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results specifically includes: Obtain industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. Operation instructions are generated based on the industrial equipment operation timestamp and the industrial equipment information to obtain original equipment operation instructions. The original equipment operation instructions are then preliminarily analyzed to obtain preliminary analysis results, which include instruction format verification, operation frequency statistics, and timing continuity statistics.

3. The device control method for the Industrial Internet of Things according to claim 1, characterized in that, The process of acquiring industrial equipment information at the industrial equipment layer, obtaining original equipment operation instructions based on the industrial equipment information, performing preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, and then further including: Based on the preliminary analysis results, obtain the equipment operation data of the industrial equipment layer, update the threshold parameters of the equipment operation data, and obtain the threshold optimization parameters. The process of updating the threshold parameters of the device operation data specifically involves: ； in, Optimize parameters for threshold. For the parameter minimization operator, For threshold parameters, As a loss indicator, For regularization weight parameters, For computational costs on edge devices, This refers to the computing capacity of edge devices.

4. The device control method for the Industrial Internet of Things according to claim 3, characterized in that, The step of performing structural analysis on the original equipment operation instructions based on the preliminary analysis results to obtain structural analysis results specifically includes: Based on the preliminary analysis results, the corresponding operation subsequences are extracted, and the operation frequency is statistically analyzed based on the subsequences. The operation frequency statistical results are then compared with a preset dynamic threshold to obtain a comparison result. If the comparison result is that the frequency value of the operation frequency statistics result is greater than or equal to the preset dynamic threshold, then the original device operation command is cyclically marked to obtain the marking result; If the comparison result is that the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, then the stack timing sequence is invoked, and the original device operation instructions are lightly screened according to the stack timing sequence to obtain the initial screening result. Based on the threshold optimization parameters, the labeling results, and the initial screening results, the structural analysis results are obtained.

5. The device control method for the Industrial Internet of Things according to claim 1, characterized in that, The semantic analysis of the structural analysis results is performed to obtain semantic analysis results, specifically as follows: The structural analysis results are subjected to data analysis to obtain structured data and unstructured data. The structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. Numerical time series analysis is performed on the structured data to obtain numerical analysis results, and text content analysis is performed on the unstructured data to obtain text analysis results. Semantic analysis results are obtained based on the numerical analysis results and the text analysis results.

6. The device control method for the Industrial Internet of Things according to claim 5, characterized in that, The numerical time series analysis of the structured data to obtain numerical analysis results specifically includes: The structured data is subjected to waveform similarity calculation to obtain waveform similarity calculation results; Numerical analysis results are obtained based on the waveform similarity calculation results; The waveform similarity calculation of the structured data specifically involves: ； in, For sequence and sequence The dynamic time-warped distance between them Align the path The minimum distance in, For sequence The first in One element, For sequence The first in Each element.

7. The device control method for the Industrial Internet of Things according to claim 5, characterized in that, The text content analysis of the unstructured data to obtain the text analysis results are as follows: The unstructured data is processed to obtain lightweight data, and vector extraction is performed on the lightweight data to obtain text vectors. The text vectors are subjected to cosine similarity calculation to obtain the cosine similarity calculation result, and the text analysis result is obtained based on the cosine similarity calculation result.

8. The device control method for the Industrial Internet of Things according to claim 4, characterized in that, The calculation of cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity specifically includes: The semantic analysis results are used to calculate similarity, and the weights of the multimodal data are configured. The multimodal data weights and the similarity results are weighted and fused to obtain a data fusion result. The data fusion result is then subject to parameter constraints to obtain cross-modal similarity.

9. The device control method for the Industrial Internet of Things according to claim 1, characterized in that, The process involves performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, performing semantic analysis on the structural analysis results to obtain semantic analysis results, and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity. The process further includes: An adaptive threshold for triggering a loop is set, and a loop decision is made based on the adaptive threshold and the cross-modal similarity according to the decision function to obtain the loop decision result. Based on the loop decision result, it is determined whether there is an abnormal operation loop in the original device operation instruction. If the loop decision result returns the first value, it is determined that there is an abnormal operation loop in the original device operation instruction; If the loop decision result returns a second value, it is determined that the original device operation instruction does not have an abnormal operation loop.

10. The device control method for the Industrial Internet of Things according to claim 9, characterized in that, The step of making iterative decisions based on the adaptive threshold and the cross-modal similarity according to the decision function is specifically as follows: ； in, Let be the decision function. For cross-modal similarity, This is an adaptive threshold.

11. The device control method for the Industrial Internet of Things according to claim 1, characterized in that, The process involves performing structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, performing semantic analysis on the structural analysis results to obtain semantic analysis results, and calculating cross-modal similarity based on the semantic analysis results to obtain cross-modal similarity. The process further includes: If the original device operation command is determined to have an abnormal operation loop based on the cross-modal similarity, then the command is generated based on the cross-modal similarity to obtain a loop termination command; The loop termination instruction is encapsulated using a protocol to obtain a target loop termination instruction, and the abnormal operation loop is terminated according to the target loop termination instruction to obtain a loop termination result.

12. The device control method for the Industrial Internet of Things according to claim 11, characterized in that, If the cross-modal similarity determines that the original device operation command does not have an abnormal operation loop, then the original device operation command is optimized based on the cross-modal similarity to obtain the target device operation command, and device control is performed based on the target device operation command, specifically including: If the original device operation command does not have an abnormal operation loop based on the cross-modal similarity, then a dynamic scheduling strategy is generated based on the cross-modal similarity. Resource reallocation is performed according to the dynamic scheduling strategy to obtain resource allocation results. The original equipment operation instructions are then optimized based on the resource allocation results to obtain target equipment operation instructions. Equipment control is then performed based on the target equipment operation instructions.

13. The device control method for the Industrial Internet of Things according to claim 12, characterized in that, The resource reallocation process based on the dynamic scheduling strategy specifically includes: ； in, To allocate new resources, For the recycling of idle resources, This is the allocation ratio coefficient. This represents the task priority gradient.

14. A device control system for the Industrial Internet of Things, characterized in that, The device control system for the Industrial Internet of Things includes: The operation instruction generation module is used to acquire industrial equipment information from the industrial equipment layer, obtain original equipment operation instructions based on the industrial equipment information, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results. The operation instruction analysis module is used to perform structural analysis on the original device operation instructions based on the preliminary analysis results to obtain structural analysis results, perform semantic analysis on the structural analysis results to obtain semantic analysis results, and perform cross-modal similarity calculation based on the semantic analysis results to obtain cross-modal similarity. The operation instruction optimization module is used to optimize the original device operation instruction according to the cross-modal similarity if it is determined that there is no abnormal operation loop in the original device operation instruction, to obtain the target device operation instruction, and to control the device according to the target device operation instruction.

15. The device control system for the Industrial Internet of Things according to claim 14, characterized in that, The operation instruction generation module includes: The data acquisition unit is used to acquire industrial equipment information and industrial equipment operation timestamps from the industrial equipment layer. The industrial equipment information includes equipment type, equipment operating status, sensor data, communication protocol, and operation log. The data analysis unit is used to generate operation instructions based on the operation timestamp of the industrial equipment and the information of the industrial equipment, obtain the original equipment operation instructions, and perform preliminary analysis on the original equipment operation instructions to obtain preliminary analysis results, wherein the preliminary analysis results include instruction format verification, operation frequency statistics and timing continuity statistics.

16. The device control system for the Industrial Internet of Things according to claim 14, characterized in that, The operation instruction analysis module includes: The data statistics unit is used to extract the corresponding operation sub-sequences based on the preliminary analysis results, perform operation frequency statistics based on the sub-sequences, obtain the operation frequency statistics results, and compare the operation frequency statistics results with a preset dynamic threshold to obtain a comparison result. A cyclic marking unit is used to cyclically mark the original device operation command if the comparison result is greater than or equal to the frequency value of the operation frequency statistics result, thereby obtaining a marking result; The data screening unit is used to call the stack timing sequence if the frequency value of the operation frequency statistics result is less than the preset dynamic threshold, and perform a lightweight screening of the original device operation instructions according to the stack timing sequence to obtain the screening result. The structural analysis unit is used to obtain structural analysis results based on the threshold optimization parameters, the labeling results, and the initial screening results; The data analysis unit is used to perform data analysis on the structural analysis results to obtain structured data and unstructured data, wherein the structured data includes vibration signal data and equipment timing data, and the unstructured data includes log text data and control command data. The text analysis unit is used to perform numerical time series analysis on the structured data to obtain numerical analysis results, and to perform text content analysis on the unstructured data to obtain text analysis results, and to obtain semantic analysis results based on the numerical analysis results and the text analysis results. A similarity calculation unit is used to calculate the similarity of the semantic analysis results, obtain the similarity results, and configure the multimodal data weights; The data fusion unit is used to perform weighted fusion of the multimodal data weights and the similarity results to obtain a data fusion result, and to impose parameter constraints on the data fusion result to obtain cross-modal similarity.

17. The device control system for the Industrial Internet of Things according to claim 14, characterized in that, The operation instruction optimization module includes: The scheduling strategy generation unit is used to generate a dynamic scheduling strategy based on the cross-modal similarity if it is determined that the original device operation instruction does not have an abnormal operation loop based on the cross-modal similarity. The resource reallocation unit is used to perform resource reallocation processing according to the dynamic scheduling strategy to obtain resource allocation results, optimize the original equipment operation instructions according to the resource allocation results to obtain target equipment operation instructions, and perform equipment control according to the target equipment operation instructions.

18. The device control system for the Industrial Internet of Things according to claim 14, characterized in that, The device control system for the Industrial Internet of Things further includes a cyclic decision-making module, which includes: The loop judgment unit is used to set an adaptive threshold for triggering the loop, perform loop decision on the adaptive threshold and the cross-modal similarity according to the decision function, obtain the loop decision result, and determine whether there is an abnormal operation loop in the original device operation instruction based on the loop decision result; The first loop determination unit is used to determine that there is an abnormal operation loop in the original device operation instruction if the loop decision result returns a first value. The second loop determination unit is used to determine that the original device operation instruction does not have an abnormal operation loop if the loop decision result returns a second value.

19. A terminal, characterized in that, The terminal includes: a memory, a processor, and a device control program for the Industrial Internet of Things (IIoT) stored in the memory and executable on the processor. When the device control program for the Industrial Internet of Things is executed by the processor, it implements the steps of the device control method for the Industrial Internet of Things as described in any one of claims 1-13.

20. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a device control program for the Industrial Internet of Things (IIoT), which, when executed by a processor, implements the steps of the device control method for the Industrial Internet of Things as described in any one of claims 1-13.

Images