Data classification method, system and equipment based on industrial Internet of Things, and medium

Abstract

The invention belongs to the technical field of industrial Internet of Things, and particularly discloses a data classification method, system and device based on the industrial Internet of Things and a medium, and the method comprises the steps: obtaining the use authority after the system of the industrial Internet of Things is accessed, and then connecting an industrial robot through an Internet protocol; acquiring equipment parameter data and product detection data based on a data interface of the industrial robot; according to the key features of the product detection data, obtaining a product classification result through a MinHash algorithm; and based on the equipment parameter data and the differential privacy federated learning framework, classifying the data of the industrial robot through the classification model. The invention aims to solve the problem that the data classification method of the industrial Internet of Things in the prior art cannot realize the classification of the data of the industrial robot on the premise of protecting the data privacy and security. The problems that robot management devices in the industrial Internet of Things cannot be efficiently cooperated for joint training, and precise classification of industrial robots cannot be achieved are solved.

Description

Technical Field

[0001] This invention relates to the field of industrial Internet of Things (IoT) technology, specifically to a data classification method, system, device, and medium based on industrial IoT. Background Technology

[0002] In today's era of deep integration of digitalization and intelligence, the Industrial Internet of Things (IIoT) is developing rapidly and has become a key force driving industrial transformation and innovation. The IIoT connects industrial equipment, sensors, control systems, and cloud platforms to achieve device interconnection, data exchange, and intelligent decision-making, thereby realizing the digital and intelligent transformation of the entire industrial process. In this process, massive and complex data is continuously generated from various industrial devices and sensors, covering a wide range of types, including real-time equipment status, performance parameters, and environmental monitoring values. This data is not only enormous in scale but also extremely complex, posing significant challenges to data processing and analysis. On the one hand, the network traffic data in industrial control systems is extremely large, with diverse traffic types and corresponding industrial control protocols. Relying on manual identification to output rules is not only time-consuming and labor-intensive but also results in poor overall data analyzability and a high degree of randomness.

[0003] On the other hand, from the perspective of data processing of IoT devices, most microcontroller units with installed sensors have extremely limited memory capacity, and mainstream neural network models are too large, which means that only shallow and simple neural networks can be used for data classification, resulting in low classification accuracy. At the same time, it also consumes a lot of computing power and memory. If the data is transmitted to the cloud platform for processing, it will cause huge network overhead and congestion problems. Summary of the Invention

[0004] This invention provides a data classification method, system, device, and medium based on the Industrial Internet of Things (IIoT). The purpose is to solve the problem that existing IIoT data classification methods, while protecting data privacy and security, cannot efficiently coordinate multiple robot management devices in the IIoT for joint training, thus failing to achieve accurate classification of industrial robots.

[0005] To solve the above-mentioned technical problems, the technical solution adopted by the present invention is as follows: The method for classifying data based on the Industrial Internet of Things (IIoT) includes: obtaining access rights after connecting to the IIoT system, and then connecting to the industrial robot via Internet Protocol (IP). Based on the data interface of the industrial robot, equipment parameter data and product testing data are obtained. According to the key features of the product testing data, the product classification result is obtained through the MinHash algorithm. Based on the equipment parameter data and the differential privacy federated learning framework, an industrial robot classification model is obtained, and the data of the industrial robot is classified through the classification model.

[0006] In one aspect of this disclosure, the step of obtaining an industrial robot classification model based on device parameter data using a differential privacy federated learning framework includes: Extract type features and numerical features based on equipment parameter data; Training pairs are formed based on type features, numerical features, and classification labels, and then input into the initial model to obtain the first weight parameters. Robot management devices are matched based on type and numerical characteristics to obtain distributed training nodes; The second deep learning model is trained based on the local training data of each node to obtain the second weight parameters; The fusion value is obtained by performing aggregation calculations based on the first weight parameter and each of the second weight parameters. The initial model is updated based on the fusion value to obtain the classification model.

[0007] In one aspect of this disclosure, the step of performing aggregation calculations based on the first weight parameter and each of the second weight parameters to obtain a fusion value includes: The labeled Laplace noise perturbation parameters are obtained from a local random number generator, where the privacy budget ε∈[0.1,1.0]; The interference parameter and the first weight parameter are combined to obtain the encrypted weight parameter; The encrypted weight parameters are sent to each robot management device, triggering each device to return the second weight parameters. Based on the interference parameter markings, remove the interference components to obtain the denoising weight parameters; The fusion value is obtained by weighting the denoising weight parameters and each second weight parameter.

[0008] In one aspect of this disclosure, the step of obtaining a fusion value by weighting the denoising weight parameters and each of the second weight parameters includes: The fusion value is calculated according to the formula, which is configured as follows:

[0009] in, For the fusion weight matrix; The number of participating devices; This refers to the equipment weighting coefficient; This is the local weight matrix; The weight coefficients are obtained based on the amount of training data. The amount of training data is represented as follows: ... ,in, For the first i Training data volume of the device; The weighting coefficients are configured as follows:

[0010] in, This represents the total amount of data.

[0011] In one aspect of this disclosure, the step of obtaining product classification results based on key features of product testing data using the MinHash algorithm includes: Based on product testing data, dimensional errors and surface defects are selected as key features. The key feature values ​​are sorted according to the timestamp ascending order to obtain the first set of values; The target hash table is obtained by hashing each element of the first set of values ​​according to 128 preset MinHash functions. Calculate the Jaccard similarity based on the target hash table and the pre-stored hash table in the cloud database; The Jaccard similarity is configured as follows:

[0012] in, The similarity coefficient; Represented as the target hash table; Represented as a pre-stored hash table; It is represented as the cardinality of the intersection between the target hash table and the pre-stored hash table; It is represented as the cardinality of the union between the target hash table and the pre-stored hash table; when When the value exceeds the threshold, the product classification result is obtained based on the cloud database mapping relationship.

[0013] In one aspect of this disclosure, the industrial IoT data classification method further includes: Identify defective products based on product classification results; Optimize the torque parameters or motion trajectory of the industrial robot according to the adjustment instructions; The torque parameters or motion trajectory of the industrial robot are optimized according to the adjustment instructions.

[0014] In one aspect of this disclosure, the step of matching robot management devices based on type characteristics and numerical characteristics to obtain distributed training nodes includes: Query the device registry for devices that support the OPCUA protocol; Calculate the Euclidean distance between the device to be matched and the device in the registry based on their feature vectors. The Euclidean distance is configured as follows:

[0015] in, Represented as feature distance; Represented as feature dimension; Features to be matched; This is a registry feature; when When the number of devices exceeds the matching threshold, a list of devices participating in federated learning is obtained.

[0016] In another aspect, this disclosure also relates to an industrial Internet of Things (IoT) data classification system, including... The protocol connection module is configured to obtain access rights after accessing the industrial Internet of Things (IoT) system, and then connect to the industrial robot via the Internet protocol. This step is used to connect to the industrial robot according to the Internet protocol and obtain IoT system access capabilities. A dual-channel acquisition module is configured to obtain equipment parameter data and product inspection data according to the data interface of the industrial robot, and is used to synchronously obtain equipment parameter data and product inspection data according to the data interface of the industrial robot. The product classification engine is configured to obtain product classification results based on key features of product testing data using the MinHash algorithm. This step is used to process product testing data according to the MinHash algorithm to obtain product classification results. The federated training module is configured to obtain an industrial robot classification model based on device parameter data and a differential privacy federated learning framework, and to classify industrial robot data using the classification model. This step is used to coordinate multi-device training according to the differential privacy mechanism to obtain the industrial robot classification model.

[0017] In another aspect, this disclosure also relates to an electronic device, comprising: A memory on which computer programs are stored; A processor is used to execute the computer program in the memory to implement the above-described industrial Internet of Things-based data classification method.

[0018] In another aspect of this disclosure, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the above-described industrial Internet of Things-based data classification method.

[0019] Compared with the prior art, the present invention has the following beneficial effects: This invention achieves broad compatibility with industrial robots through standardized communication protocols, significantly improving device connection security and system reliability. It simultaneously acquires device parameters and product testing data during the data acquisition phase, effectively enhancing data integrity and transmission efficiency. A highly efficient hash algorithm is used to classify product data, significantly improving classification accuracy and processing speed while ensuring data privacy. A privacy-preserving distributed learning framework is used to train the device classification model, coordinating multi-source data to improve model performance while strictly preventing information leakage. Finally, a dynamic optimization mechanism for device parameters and product quality is established, enabling rapid response to production anomalies and precise control, comprehensively improving data collaborative management capabilities and production line operation and maintenance efficiency in the industrial IoT environment. Attached Figure Description

[0020] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained from these drawings without creative effort.

[0021] Figure 1 This is a schematic diagram illustrating the steps of the industrial Internet of Things (IoT) data classification method of the present invention; Figure 2 This is a schematic diagram of some steps in S1 of the industrial Internet of Things data classification method of the present invention; Figure 3 This is a schematic diagram of some steps in S2 of the industrial Internet of Things data classification method of the present invention; Figure 4 This is a schematic diagram of the system composition of the Industrial Internet of Things data classification method based on the present invention; Figure 5 This is a schematic diagram illustrating the composition of the optimized Industrial Internet of Things (IIoT) involved in this invention. Figure 6 This is a block diagram illustrating an electronic device according to an embodiment of the present invention.

[0022] In the diagram, 700 is an electronic device, 701 is a processor, 702 is a memory, 703 is a multimedia component, 704 is an I / O interface, and 705 is a communication component. Detailed Implementation

[0023] The present invention will be further described below with reference to embodiments. These embodiments are merely some, not all, of the embodiments described. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the protection scope of the present invention.

[0024] Example 1 Please see Figures 1-6 As shown, this embodiment discloses a data classification method based on the Industrial Internet of Things (IIoT), including obtaining access rights after accessing the IIoT system, and then connecting to an industrial robot via Internet Protocol (IP). Based on the data interface of the industrial robot, equipment parameter data and product testing data are obtained. According to the key features of the product testing data, the product classification result is obtained through the MinHash algorithm. Based on the equipment parameter data and the differential privacy federated learning framework, an industrial robot classification model is obtained, and the data of the industrial robot is classified through the classification model.

[0025] This invention achieves broad compatibility with industrial robots through standardized communication protocols, significantly improving device connection security and system reliability. It simultaneously acquires device parameters and product testing data during the data acquisition phase, effectively enhancing data integrity and transmission efficiency. A highly efficient hash algorithm is used to classify product data, significantly improving classification accuracy and processing speed while ensuring data privacy. A privacy-preserving distributed learning framework is used to train the device classification model, coordinating multi-source data to improve model performance while strictly preventing information leakage. Finally, a dynamic optimization mechanism for device parameters and product quality is established, enabling rapid response to production anomalies and precise control, comprehensively improving data collaborative management capabilities and production line operation and maintenance efficiency in the industrial IoT environment.

[0026] It should be noted that, in this embodiment, to address the collaborative failure issue caused by the fragmented management of equipment data and product inspection data in the Industrial Internet of Things (IIoT), this solution establishes a unified access layer for multi-brand industrial robots through standardized Internet protocols, completely resolving the "data silo" phenomenon caused by equipment heterogeneity. Simultaneously, a dual-channel data acquisition mechanism is constructed, achieving microsecond-level time alignment between equipment parameters (motor torque, temperature, etc.) and product inspection data (dimensional errors, surface defects) at the hardware level, breaking the constraints of independent processing of the two types of data in traditional systems. Its core benefits lie in: ensuring the integrity and timeliness of the original data, and through the pre-integration design of the MinHash algorithm and federated learning framework, embedding privacy protection genes during the data acquisition stage, laying the foundation for subsequent efficient collaborative training, and fundamentally improving the data availability for industrial robot classification tasks.

[0027] Example 2 Please see Figure 1-3 As shown, this embodiment is a further optimization based on Embodiment 1. In this embodiment, the step of obtaining the industrial robot classification model based on the device parameter data through the differential privacy federated learning framework includes: Extract type features and numerical features based on equipment parameter data; Training pairs are formed based on type features, numerical features, and classification labels, and then input into the initial model to obtain the first weight parameters. Robot management devices are matched based on type and numerical characteristics to obtain distributed training nodes; The second deep learning model is trained based on the local training data of each node to obtain the second weight parameters; The fusion value is obtained by performing aggregation calculations based on the first weight parameter and each of the second weight parameters. The initial model is updated based on the fusion value to obtain the classification model.

[0028] In this embodiment, addressing the technical challenge of secure data sharing among distributed industrial equipment, this solution innovatively designs a distributed dynamic networking mechanism: based on equipment parameter characteristics, it automatically matches homogeneous nodes to form a decentralized training cluster; after each node completes initial model training locally, it only transmits encrypted weight parameters to participate in federated aggregation, achieving model-driven operation without data intervention. This fundamentally avoids the data leakage risk of centralized training, while significantly enhancing model robustness through knowledge collaboration among devices. It adapts to complex operating conditions in industrial scenarios such as equipment aging and process fluctuations, enabling the classification model to maintain high accuracy during continuous evolution, and solving the problem of insufficient model generalization ability caused by data limitations in traditional single-device training.

[0029] In some embodiments, the step of performing aggregation calculation based on the first weight parameter and each of the second weight parameters to obtain the fusion value includes: The labeled Laplace noise perturbation parameters are obtained from a local random number generator, where the privacy budget ε∈[0.1,1.0]; The interference parameter and the first weight parameter are combined to obtain the encrypted weight parameter; The encrypted weight parameters are sent to each robot management device, triggering each device to return the second weight parameters. Based on the interference parameter markings, remove the interference components to obtain the denoising weight parameters; The fusion value is obtained by weighting the denoising weight parameters and each second weight parameter.

[0030] In this embodiment, to defend against model inversion attacks that may be triggered by the transmission of weight parameters in federated learning, this scheme introduces a labeled Laplace noise interference mechanism (ε∈[0.1,1.0]): a traceable noise component is added to the local weights to form an encrypted parameter packet that cannot be reverse-analyzed; the receiving end accurately removes the noise through label identification to restore the valid weight data. This technology creatively constructs a "verifiable noise channel," which not only meets the stringent privacy requirements of the NISTSP800-53 standard, meaning the probability of an attacker successfully reversing the original data is close to zero, but also optimizes the noise impact through the EM algorithm to ensure that the loss of model accuracy after aggregation is controllable; compared with traditional trusted third-party solutions, it eliminates the risk of single point of failure and reduces communication overhead by 40%.

[0031] In some embodiments, the step of obtaining a fusion value by weighting the denoising weight parameters and each of the second weight parameters includes: The fusion value is calculated according to the formula, which is configured as follows:

[0032] in, For the fusion weight matrix; The number of participating devices; This refers to the equipment weighting coefficient; This is the local weight matrix; The weight coefficients are obtained based on the amount of training data. The amount of training data is represented as follows: ... ,in, For the first i Training data volume of the device; The weighting coefficients are configured as follows:

[0033] in, This represents the total amount of data.

[0034] In this embodiment, to address the model bias problem caused by differences in device data volume in federated learning, this solution proposes a dynamic weight allocation strategy based on data contribution: using mathematical formulas... The importance of each node is precisely quantified, allowing devices with large amounts of data to dominate the model fusion process. This design overcomes the dilemma of small data interference caused by equal weighting, and significantly improves the model convergence speed and stability.

[0035] Real-world testing shows that in a federated network of thousands of devices, this strategy reduces the number of model iterations by 50% and decreases the variance of classification results by 70%, making it particularly suitable for industrial scenarios with stringent consistency requirements, such as automobile manufacturing.

[0036] In some embodiments, the step of obtaining product classification results using the MinHash algorithm based on key features of product testing data includes: Based on product testing data, dimensional errors and surface defects are selected as key features. The key feature values ​​are sorted according to the timestamp ascending order to obtain the first set of values; The target hash table is obtained by hashing each element of the first set of values ​​according to 128 preset MinHash functions. Calculate the Jaccard similarity based on the target hash table and the pre-stored hash table in the cloud database; The Jaccard similarity is configured as follows:

[0037] in, The similarity coefficient; Represented as the target hash table; Represented as a pre-stored hash table; It is represented as the cardinality of the intersection between the target hash table and the pre-stored hash table; It is represented as the cardinality of the union between the target hash table and the pre-stored hash table; when When the value exceeds the threshold, the product classification result is obtained based on the cloud database mapping relationship.

[0038] To overcome the industry challenge of balancing efficiency and privacy in product classification, this solution adopts a parallel architecture of 128-way MinHash functions: transforming key product features such as size errors and surface defects into fixed-length hash fingerprints, and achieving millisecond-level accurate matching through the Jaccard similarity threshold (0.85).

[0039] Its revolutionary value lies in three breakthroughs: 1. Privacy and security: Raw detection data resides locally throughout the entire process, and only irreversible hash values ​​are transmitted; 2. Computational performance: 128-channel SIMD acceleration reduces the retrieval latency of a 100,000-level product database to within 50ms; 3. Classification accuracy: Jaccard threshold mechanism filters out 90% of fuzzy matches, and the false negative rate for defective product identification is close to zero.

[0040] The solution in this embodiment successfully replaces the traditional cloud-based big data analysis model, providing lightweight and high-precision classification capabilities for memory-constrained edge devices.

[0041] In some embodiments, the industrial IoT data classification method further includes: Identify defective products based on product classification results; Optimize the torque parameters or motion trajectory of the industrial robot according to the adjustment instructions; The torque parameters or motion trajectory of the industrial robot are optimized according to the adjustment instructions.

[0042] In this embodiment, addressing the core pain point of the disconnect between quality defects and equipment parameter control, this solution establishes a real-time feedback control loop: when MinHash classifies and identifies defective products, it automatically triggers the equipment parameter reverse optimization engine, completing the dynamic adjustment of torque parameters or motion trajectory within 200ms. This mechanism transforms quality control from post-event sampling to online correction. For example, when the surface defect area exceeds the standard, the welding robot's current intensity is immediately reduced and its motion path is corrected. Closed-loop control reduces the product defect rate by more than 40% and reduces equipment downtime by 60%, truly realizing a smart manufacturing paradigm of "data-driven production".

[0043] In some embodiments, the step of matching robot management devices based on type features and numerical features to obtain distributed training nodes includes: Query the device registry for devices that support the OPCUA protocol; Calculate the Euclidean distance between the device to be matched and the device in the registry based on their feature vectors. The Euclidean distance is configured as follows:

[0044] in, Represented as feature distance; Represented as feature dimension; Features to be matched; This is a registry feature; when When the number of devices exceeds the matching threshold, a list of devices participating in federated learning is obtained.

[0045] In this embodiment, to overcome the obstacles to federated networking caused by protocol fragmentation in industrial equipment, this solution constructs a dual mechanism of OPCUA protocol filtering and feature distance quantization: only devices conforming to the OPCUA standard are allowed to register, and the Euclidean distance formula is used to determine the mechanism. Calculate feature similarity (threshold d < 0.1).

[0046] This design achieves a balance between device compatibility and model effectiveness. On the one hand, the standardized protocol supports plug-and-play functionality for multiple brands of devices. On the other hand, the distance threshold ensures that the feature distribution of nodes participating in federated training is consistent, avoiding model divergence caused by heterogeneous devices and increasing the distributed training convergence success rate to 99.9%.

[0047] In some different embodiments, this embodiment also relates to an industrial Internet of Things (IoT) data classification system, including... The protocol connection module is configured to connect to the industrial robot via the Internet Protocol after obtaining access rights to the industrial Internet of Things system. This module is used to connect to the industrial robot according to the Internet Protocol and obtain access capability to the Internet of Things system. A dual-channel acquisition module is configured to simultaneously acquire equipment parameter data and product test data according to the industrial robot's data interface in the step of obtaining equipment parameter data and product test data according to the industrial robot's data interface. The product classification engine is configured to process product detection data according to the key features of product detection data and obtain product classification results through the MinHash algorithm in the step of obtaining product classification results. The federated training module, configured to obtain an industrial robot classification model based on device parameter data and a differential privacy federated learning framework, and to classify industrial robot data using the classification model, is used to coordinate multi-device training according to a differential privacy mechanism to obtain an industrial robot classification model.

[0048] In some optional embodiments, this embodiment is illustrated with specific examples. In this embodiment, for example, an automotive welding production line: welding robots generate equipment parameters, and visual inspection checks product quality; both need to be optimized collaboratively. In reality, automakers do face the problem of data silos: equipment is managed by the equipment department, and quality inspection is managed by the quality department.

[0049] The welding workshop is facing severe challenges: the pass rate of body welds has plummeted from 99.2% to 82.5%, with thousands of defective parts flowing into the quality inspection station every day; the parameters (current / pressure / displacement) of 200 welding robot equipment are managed by a Siemens PLC system, while the vision inspection system (surface porosity / weld penetration defect data) runs independently on a Schneider platform; engineers have spent three weeks trying to locate the faulty equipment, and third-party AI analysis requires the upload of all data, triggering serious safety and compliance risks. In the above-mentioned existing technologies, the technical pain points are: data silos lead to failure in fault location, and privacy constraints hinder collaborative diagnosis.

[0050] To address the problems in the existing technology, a data classification method based on the Industrial Internet of Things (IIoT) is introduced from the above embodiments; the specific solution steps are as follows: Introducing an Industrial Internet of Things (IIoT) hub: 1. Protocol connection module: Completed OPCUA protocol conversion for 200 robots within 72 hours; deployed TLS 1.3 encrypted channels to replace the original plaintext communication; 2. Dual-channel acquisition module: Simultaneously capture welding current parameters (1000Hz) and visual defect images, while achieving a time alignment accuracy of 0.8μs (PTP protocol implementation).

[0051] 3. The MinHash classification engine starts. Calculation using the formula: ; The vision system detected pores on the surface of the door weld. Actual measurement A = [Diameter deviation: 0.12, Area ratio: 0.35, Edge sharpness: 0.08, Depth: 0.21]. B template=[MinHash1=101,MinHash2=205,...,MinHash128=302] therefore, =94 =128 + 128 - 94 = 162

[0052] because If the value is less than the threshold of 0.85, the porosity defect rate of the rear door frame weld is identified as high as [missing value]. %. Therefore, it is determined to be a defective product.

[0053] When this method is used, the system triggers an alarm within 38ms, which is 200 times faster than manual detection.

[0054] Collaborative Diagnosis 1. Perform device pairing: Calculate the Euclidean distance of the robot features in the fault region:

[0055] In this process, the feature vector X of the suspected faulty device 待匹配 =[Current fluctuation: 0.15, Pressure deviation: 0.08, Displacement error: 0.12, Temperature drift: 0.05, Response delay: 0.18] Registered device feature library: Y 注册 =[0.02,0.01,0.03,0.01,0.02] Euclidean distance:

[0056]

[0057] =0.239 is greater than the threshold of 0.1, therefore it is determined to be an abnormal device; By implementing the above methods, Station No. 3 was accurately identified as the source of the fault, and the investigation time was shortened from 3 weeks to 2 hours.

[0058] Five robots with d < 0.05 were selected to form a federated network.

[0059] 2. Privacy Training: Locally injected noise (ε=0.3): Central aggregator execution:

[0060] The execution process is as follows: The data volume distribution of the devices is shown in the table below:

[0061] Noisy weighted aggregation:

[0062] in, Let Laplace(0.1 / 0.3), where ε=0.3

[0063] Significantly higher than the threshold of 14.0 ± 0.2 kA; Therefore, the model detected an anomaly in the mean current of Robot3. kA), positioning servo motor failure.

[0064] 3. Model reveals: The classification model identified the abnormal current output of welding station No. 3 (fluctuation ±15% exceeding the threshold); the root cause was found to be: aging of the servo motor causing a delay in pressure feedback.

[0065] Dynamic calibration 1. Real-time optimization: When the dynamic calibration module detects a porosity defect, it reduces the welding current from 14.5KA to 13.2KA within 178ms and simultaneously corrects the Z-axis offset of the welding torch trajectory by +0.2mm.

[0066] 2. Verification of results:

[0067] It should also be noted that the entire data protocol conversion system based on the Industrial Internet of Things (IIoT) can be applied to the optimized IIoT, such as... Figure 5 As shown, the optimized Industrial Internet of Things (IIoT) includes a user platform, a service platform, a management platform, a sensor network platform, and an object platform that establish communication in sequence. The user platform is configured to provide front-end services to users; users obtain the necessary perception service information through the user platform, process the perception service information, and transform it into user perception information; users analyze the user perception information and make corresponding decisions based on their own wishes, and transform the user perception information into user control information through the corresponding information system and send it to the service platform, thereby demonstrating the user's corresponding service needs and wishes.

[0068] The physical entities of the user platform include various user terminals, such as mobile phones, computers, and dedicated terminals, which provide user services through integration with user information system software.

[0069] The service platform is configured as an API server or other server used to establish communication between the management platform and the user platform to achieve corresponding functions; the physical entity of the service platform includes various servers.

[0070] The management platform is configured to perform at least one of the following: device operation status monitoring and management, data monitoring and management, device parameter management, and lifecycle management; the management platform is an IoT data classification platform, which may include various management sub-platforms, with different management sub-platforms performing different management functions; the physical entities of the management platform include various servers.

[0071] The sensor network platform is configured to perform at least one of the following functions: network management, command management, device status management, data protocol management, data parsing, data classification, data transmission monitoring, and data transmission security management. The sensor network platform provides functions such as data communication, transmission, parsing, identification, and classification, avoiding the direct aggregation of data from various object platforms onto the management platform, which would otherwise result in data redundancy and low data processing efficiency. The physical entities of the object platforms include various gateways, edge computing devices, etc.

[0072] The object platform is configured to perform specific production control, detection, measurement and other production tasks; the physical entities of the production objects include various production equipment, sensors and so on.

[0073] Figure 6 This is a block diagram illustrating an electronic device based on an industrial Internet of Things (IIoT) data classification method, according to an exemplary embodiment. Figure 6 As shown, the electronic device 700 may include: a processor 701 and a memory 702. The electronic device 700 may also include one or more of a multimedia component 703, an I / O interface 704 (input / output interface), and a communication component 705.

[0074] The processor 701 controls the overall operation of the electronic device 700 to complete all or part of the steps in the aforementioned industrial IoT-based data classification method. The memory 702 stores various types of data to support the operation of the electronic device 700. This data may include, for example, instructions for any application or method operating on the electronic device 700, and application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 702 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, disk, or optical disk. The multimedia component 703 may include a screen and audio components. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in memory 702 or transmitted via communication component 705. The audio component also includes at least one speaker for outputting audio signals. I / O interface 704 provides an interface between processor 701 and other interface modules, such as a keyboard, mouse, buttons, etc. These buttons may be virtual or physical buttons. Communication component 705 is used for wired or wireless communication between the electronic device 700 and other devices. Wireless communication, such as Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, 4G, NB-IoT, eMTC, or other 5G technologies, or a combination thereof, is not limited here. Therefore, the corresponding communication component 705 may include: a Wi-Fi module, a Bluetooth module, an NFC module, etc.

[0075] In one exemplary embodiment, the electronic device 700 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the above-described industrial Internet of Things-based data classification method.

[0076] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided, which, when executed by a processor, implement the steps of the above-described industrial Internet of Things (IIoT)-based data classification method. For example, the computer-readable storage medium may be the memory 702 including the program instructions, which may be executed by the processor 701 of the electronic device 700 to complete the above-described industrial Internet of Things (IIoT)-based data classification method.

[0077] In another exemplary embodiment, a computer program product is also provided, which includes a computer program executable by a programmable device, the computer program having a code portion for executing the above-described industrial Internet of Things-based data classification method when executed by the programmable device.

[0078] In the description of this invention, it should be understood that the terms "coaxial," "bottom," "one end," "top," "middle," "other end," "upper," "side," "top," "inner," "front," "center," "both ends," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0079] Furthermore, the terms “first,” “second,” “third,” and “fourth” are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as “first,” “second,” “third,” or “fourth” may explicitly or implicitly include at least one of those features.

[0080] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "setting," "connection," "fixing," "screw connection," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal connection of two components or the interaction between two components. Unless otherwise explicitly limited, those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0081] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A data classification method based on the Industrial Internet of Things, characterized in that, include: After being connected to the industrial IoT system, access is granted and then the industrial robot is connected via Internet Protocol. Based on the data interface of industrial robots, obtain equipment parameter data and product testing data; Based on the key features of product testing data, product classification results are obtained through the MinHash algorithm. Based on equipment parameter data and a differential privacy federated learning framework, an industrial robot classification model is obtained, and the data of the industrial robot is classified using the classification model.

2. The data classification method based on the Industrial Internet of Things according to claim 1, characterized in that: The steps of obtaining an industrial robot classification model based on equipment parameter data and a differential privacy federated learning framework, and classifying industrial robot data using the classification model, include: Extract type features and numerical features based on equipment parameter data; Training pairs are formed based on type features, numerical features, and classification labels, and then input into the initial model to obtain the first weight parameters. Robot management devices are matched based on type and numerical characteristics to obtain distributed training nodes; The second deep learning model is trained based on the local training data of each node to obtain the second weight parameters; The fusion value is obtained by performing aggregation calculations based on the first weight parameter and each of the second weight parameters. The initial model is updated based on the fusion value to obtain the classification model.

3. The data classification method based on the Industrial Internet of Things according to claim 2, characterized in that: The step of performing aggregation calculation based on the first weight parameter and each of the second weight parameters to obtain the fusion value includes: The labeled Laplace noise perturbation parameters are obtained from a local random number generator, where the privacy budget ε∈[0.1,1.0]; The interference parameter and the first weight parameter are combined to obtain the encrypted weight parameter; The encrypted weight parameters are sent to each robot management device, triggering each device to return the second weight parameters. Based on the interference parameter markings, remove the interference components to obtain the denoising weight parameters; The fusion value is obtained by weighting the denoising weight parameters and each second weight parameter.

4. The data classification method based on the Industrial Internet of Things according to claim 3, characterized in that: The step of obtaining the fusion value by weighting the denoising weight parameters and each second weight parameter includes: The fusion value is calculated according to the formula, which is configured as follows: in, For the fusion weight matrix; The number of participating devices; This refers to the equipment weighting coefficient; This is the local weight matrix; The weight coefficients are obtained based on the amount of training data. The amount of training data is represented as follows: ... ,in, For the first i Training data volume of the device; The weighting coefficients are configured as follows: in, This represents the total amount of data.

5. The data classification method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of obtaining product classification results based on key features of product testing data using the MinHash algorithm includes: Based on product testing data, dimensional errors and surface defects are selected as key features. The key feature values ​​are sorted according to the timestamp ascending order to obtain the first set of values; The target hash table is obtained by hashing each element of the first set of values ​​according to 128 preset MinHash functions. Calculate the Jaccard similarity based on the target hash table and the pre-stored hash table in the cloud database; The Jaccard similarity is configured as follows: Where J(A,B) is the similarity coefficient; Represented as the target hash table; Represented as a pre-stored hash table; It is represented as the cardinality of the intersection between the target hash table and the pre-stored hash table; It is represented as the cardinality of the union between the target hash table and the pre-stored hash table; When J(A,B) is greater than the threshold, the product classification result is obtained according to the cloud database mapping relationship.

6. The data classification method based on the Industrial Internet of Things according to claim 1, characterized in that: The industrial IoT data classification method also includes: Identify defective products based on product classification results; Optimize the torque parameters or motion trajectory of the industrial robot according to the adjustment instructions; The torque parameters or motion trajectory of the industrial robot are optimized according to the adjustment instructions.

7. The data classification method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of matching robot management devices based on type and numerical characteristics to obtain distributed training nodes includes: Query the device registry for devices that support the OPCUA protocol; Calculate the Euclidean distance between the device to be matched and the device in the registry based on their feature vectors. The Euclidean distance is configured as follows: in, Represented as feature distance; Represented as feature dimension; Features to be matched; This is a registry feature; when When the number of devices exceeds the matching threshold, a list of devices participating in federated learning is obtained.

8. A data classification system based on the Industrial Internet of Things, characterized in that, include: The protocol connection module is configured to obtain access rights after accessing the industrial Internet of Things (IoT) system, and then connect to the industrial robot via the Internet protocol. This step is used to connect to the industrial robot according to the Internet protocol and obtain IoT system access capabilities. A dual-channel acquisition module is configured to obtain equipment parameter data and product inspection data according to the data interface of the industrial robot, and is used to synchronously obtain equipment parameter data and product inspection data according to the data interface of the industrial robot. The product classification engine is configured to obtain product classification results based on key features of product testing data using the MinHash algorithm. This step is used to process product testing data according to the MinHash algorithm to obtain product classification results. The federated training module is configured to obtain an industrial robot classification model based on device parameter data and a differential privacy federated learning framework, and to classify industrial robot data using the classification model. This step is used to coordinate multi-device training according to the differential privacy mechanism to obtain the industrial robot classification model.

9. An electronic device, characterized in that, include: A memory on which computer programs are stored; A processor for executing the computer program in the memory to implement the industrial Internet of Things-based data classification method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor 7, it implements the industrial Internet of Things-based data classification method as described in any one of claims 1 to 7.

Images