Data protocol conversion method and system based on industrial Internet of Things

Abstract

The invention belongs to the technical field of industrial Internet of Things data processing, and particularly discloses a data protocol conversion method and system based on the industrial Internet of Things, and the method comprises the steps: extracting a feature vector through a protocol feature recognition model according to an original data stream of target equipment, and obtaining a matching result of a dynamic protocol fingerprint database; according to a matching result of the dynamic protocol fingerprint database, judging whether a protocol learning mode is triggered to obtain a newly added fingerprint and an analysis rule; according to the analyzed structured source data and the domain ontology library, semantic annotation is carried out through a semantic understanding engine, and semantic data with physical quantity tags are obtained; and according to the real-time resource state parameters, a conversion strategy is dynamically selected, an unloading point is decided and calculated, and a lightweight and full-amount depth conversion instruction is obtained. The invention aims to solve the problem of low protocol conversion efficiency caused by protocol isomerism, semantic difference and resource constraint in the industrial Internet of Things in the prior art.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of industrial Internet of Things data processing, and particularly relates to a data protocol conversion method and system based on an industrial Internet of Things. BACKGROUND

[0002] Under the background of rapid development of the industrial Internet of Things (IIoT), various devices and systems in an industrial field present a highly heterogeneous characteristic. Sensors, controllers, actuators and host computer systems of different manufacturers often adopt exclusive data communication protocols, such as Modbus, Profinet, OPC UA, EtherCAT and the like. These protocols have significant differences in data formats, transmission mechanisms, communication rates and the like, which leads to the problem of "information island" in data interaction between devices. With the deepening of application requirements such as intelligent manufacturing and industrial big data analysis, the industrial field has higher requirements for real-time data collection, integration and sharing across devices and systems. The traditional protocol adaptation mode depending on customized interface development or manual configuration not only has a long development cycle and high maintenance cost, but also is difficult to meet the scene requirements of large-scale device access and dynamic protocol updating. In addition, the high reliability, low delay and strong security requirements of the industrial environment further aggravate the technical challenges in the protocol conversion process. How to realize efficient, stable and secure conversion between different protocols has become a key technical bottleneck for connecting industrial data links and supporting deep application of the industrial Internet of Things. SUMMARY

[0003] The application provides a data protocol conversion method and system based on an industrial Internet of Things, and aims to solve the problem of low protocol conversion efficiency caused by protocol heterogeneity, semantic difference and resource constraint in the industrial Internet of Things in the prior art.

[0004] To solve the above technical problems, the technical scheme adopted by the application is as follows: The data protocol conversion method based on the industrial Internet of Things comprises the following steps: according to original data flow of a target device, a feature vector is extracted through a protocol feature recognition model to obtain a matching result of a dynamic protocol fingerprint library; according to the matching result of the dynamic protocol fingerprint library, it is judged whether a protocol learning mode is triggered to obtain a new fingerprint and a parsing rule; according to the parsed structured source data and a domain ontology library, semantic labeling is performed through a semantic understanding engine to obtain semantic data with a physical quantity label; according to real-time resource state parameters, a conversion strategy is dynamically selected and a calculation offloading point is decided to obtain lightweight and full-depth conversion instructions; and according to the semantic mapping result and the conversion strategy, data is encapsulated according to a target protocol and is transmitted through a priority queue to obtain feedback optimization information of a target system.

[0005] In an aspect of the present disclosure, the step of obtaining a matching result of a dynamic protocol fingerprint library according to a feature vector extracted by a protocol feature recognition model from a raw data stream of a target device comprises: According to the actively sent probe instruction set, the interactive response data stream of the unknown protocol is obtained; According to the frame structure features and check mode of the response data stream, a protocol feature template is generated; According to the pre-set rule generator, a register mapping relationship is derived to obtain executable analysis rules.

[0006] In an aspect of the present disclosure, the step of obtaining semantic data with physical quantity labels by semantic annotation by a semantic understanding engine according to the parsed structured source data and the domain ontology library comprises: According to the device ID and the register address, the domain ontology library is queried to obtain physical quantity semantic labels and units; According to the semantic description of the target system field, the source data points are matched by cosine similarity calculation to obtain automatically generated conversion logic chains.

[0007] In an aspect of the present disclosure, the step of obtaining automatically generated conversion logic chains according to the semantic description of the target system field by cosine similarity calculation comprises: According to the difference between the source data units and the target units, a dimensional conversion formula library is called to obtain standard unit data; According to the state code mapping table, the binary state value is converted into a text description readable by the target system.

[0008] In an aspect of the present disclosure, the step of obtaining semantic data with physical quantity labels by semantic annotation by a semantic understanding engine according to the parsed structured source data and the domain ontology library comprises: According to the result that the CPU utilization exceeds the threshold value, a lightweight conversion mode is switched to obtain simplified data that only retains key fields; According to the memory occupancy rate and the task calculation complexity evaluation, the aggregation calculation task is offloaded to the cloud to obtain a cross-node cooperative processing state code.

[0009] In an aspect of the present disclosure, the step of obtaining simplified data that only retains key fields by switching to a lightweight conversion mode according to the result that the CPU utilization exceeds the threshold value comprises: According to the data priority label, low importance data points are filtered to obtain a transmission data set with a compression rate ≥ 60%.

[0010] In an aspect of the present disclosure, the step of obtaining semantic data with physical quantity labels by semantic annotation by a semantic understanding engine according to the parsed structured source data and the domain ontology library further comprises: According to the semantic label conflict result, the ontology library mapping relationship is corrected through a manual annotation interface, and a conversion rule after conflict resolution is obtained.

[0011] In an aspect of the present disclosure, the step of performing semantic labeling through a semantic understanding engine according to the parsed structured source data and the domain ontology library to obtain semantic data with physical quantity labels further includes: According to network bandwidth fluctuation data, any one of Protobuf and JSON is selected as an encapsulation format to obtain a serialized message with an adaptive compression ratio.

[0012] In an aspect of the present disclosure, the step of encapsulating data according to the semantic mapping result and the conversion strategy, and transmitting the data through a priority queue to obtain feedback optimization information of a target system includes: According to target system error code feedback, a defect of a protocol fingerprint library analysis rule is located, and an iteratively optimized feature template is obtained. According to semantic mapping failure records, the correlation weight of the domain ontology library is adjusted to obtain an enhanced semantic matching accuracy.

[0013] In another aspect of the present disclosure, the present disclosure also relates to a data protocol conversion system based on an industrial Internet of Things, which includes a protocol fingerprint learning module, a semantic understanding engine module, a resource scheduling module, and a protocol encapsulation and transmission module. The protocol fingerprint learning module is configured to obtain a dynamic protocol fingerprint library by extracting a feature vector through a protocol feature recognition model according to a target device raw data stream, and obtain a matching result of the dynamic protocol fingerprint library. The semantic understanding engine module is configured to obtain semantic data with physical quantity labels by performing semantic labeling through a semantic understanding engine according to parsed structured source data and a domain ontology library, and obtain semantic data and a conversion rule. The resource scheduling module is configured to obtain adaptive strategies and unloading instructions by dynamically selecting a conversion strategy and deciding a calculation unloading point according to real-time resource state parameters, and obtain lightweight and full-depth conversion instructions. The protocol encapsulation and transmission module is configured to obtain a target system compatible transmission result by dynamically selecting a conversion strategy and deciding a calculation unloading point according to real-time resource state parameters, and obtaining lightweight and full-depth conversion instructions.

[0014] Compared with the prior art, the present application has the following beneficial effects: The application solves the problem of relying on manual protocol analysis in traditional solutions by constructing a four-dimensional collaborative mechanism of protocol self-learning, semantic intelligent analysis, resource dynamic scheduling and closed-loop feedback optimization, dynamically matching and self-learning through a dynamic protocol fingerprint library, shortening the new device access period from a week to an hour, and supporting automatic expansion of private protocols. At the same time, based on the semantic annotation and rule chain generation of the domain ontology library, the error risk of manual configuration of the mapping table is eliminated, the accuracy of cross-system data fusion is improved, and the conversion failure caused by semantic ambiguity is completely solved. According to the real-time load, the lightweight / full-depth mode and the calculation offloading strategy are dynamically switched, the resource bottleneck of the edge node is broken through, the resource consumption is reduced and the zero packet loss rate is maintained in the high-load scene. Finally, the error code feedback from the target system is used to drive the iteration of the protocol fingerprint library and the semantic mapping rule, which reduces the system maintenance cost and continuously improves the conversion accuracy. The advantages of such a setting are that the multi-source heterogeneous device access efficiency can be significantly improved, the bottleneck of high dependence on manual configuration in the protocol analysis of traditional solutions can be completely eliminated, the data semantic mapping accuracy can be improved, the data fusion obstacles caused by cross-system semantic conflicts can be solved, and the resource constraint limitation in the edge computing scene can be broken through to ensure the system stability in high-load working conditions, and the continuous self-optimization capability is established, which greatly reduces the whole life cycle operation and maintenance complexity. BRIEF DESCRIPTION OF DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments. It should be understood that the following drawings only show some embodiments of the application, and therefore should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can also be obtained without creative labor.

[0016] Figure 1 The steps of the data protocol conversion method based on industrial Internet of Things of the application are shown in the figure. Figure 2 The partial steps of S1 in the data protocol conversion method based on industrial Internet of Things of the application are shown in the figure. Figure 3 The partial steps of S2 in the data protocol conversion method based on industrial Internet of Things of the application are shown in the figure. Figure 4 The composition of the data protocol conversion method system based on industrial Internet of Things of the application is shown in the figure. Figure 5 The composition of the optimized industrial Internet of Things related to the application is shown in the figure. Figure 6 The block diagram of an electronic device according to an embodiment of the application is shown in the figure.

[0017] In the figure, 700-electronic device, 701-processor, 702-memory, 703-multimedia component, 704-I / O interface, 705-communication component. DETAILED DESCRIPTION

[0018] The application will be further described below in conjunction with the embodiments, and the described embodiments are only part of the embodiments of the application, not all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor are within the protection scope of the application.

[0019] Embodiment one Please refer to Figures 1-6 As shown in the figure, the embodiment discloses a data protocol conversion method based on industrial Internet of Things, which comprises the following steps: according to the original data stream of a target device, a feature vector is extracted through a protocol feature recognition model to obtain a matching result of a dynamic protocol fingerprint library; according to the matching result of the dynamic protocol fingerprint library, it is judged whether a protocol learning mode is triggered to obtain a new fingerprint and a parsing rule; according to the parsed structured source data and a domain ontology library, semantic annotation is carried out through a semantic understanding engine to obtain semantic data with physical quantity labels; according to real-time resource state parameters, a conversion strategy is dynamically selected and a calculation offloading point is decided to obtain lightweight, full-depth conversion instructions; according to the semantic mapping result and the conversion strategy, data is encapsulated according to a target protocol and transmitted through a priority queue to obtain feedback optimization information of a target system.

[0020] The application solves the problem that the traditional scheme relies on manual protocol analysis by constructing a four-dimensional collaborative mechanism of protocol self-learning, semantic intelligent parsing, resource dynamic scheduling and closed-loop feedback optimization and a dynamic protocol fingerprint library matching and self-learning mechanism, shortens the new device access period from a week level to an hour level, and supports automatic expansion of private protocols; at the same time, the semantic annotation and rule chain generation based on the domain ontology library eliminate the error risk of manual configuration of the mapping table, improve the cross-system data fusion accuracy, and completely solve the conversion failure caused by semantic ambiguity; according to the real-time load, the lightweight / full-depth mode and the calculation offloading strategy are dynamically switched, the resource bottleneck of the edge node is broken through, the resource consumption is reduced and the zero packet loss rate is maintained in the high-load scene; finally, the error code fed back by the target system is used to drive the iteration of the protocol fingerprint library and the semantic mapping rule, the system maintenance cost is reduced, and the continuous evolution of the conversion accuracy is realized. The advantage of such setting is that the multi-source heterogeneous device access efficiency can be significantly improved, the bottleneck of high dependence on manual configuration of protocol analysis in the traditional scheme is completely eliminated, the data semantic mapping precision is improved, the data fusion obstacle caused by cross-system semantic conflict is solved, and the resource constraint limitation in the edge computing scene is broken through, the system stability under high-load working conditions is ensured, the continuous self-optimization capability is simultaneously established, and the whole life cycle operation and maintenance complexity is greatly reduced.

[0021] The specific workflow of the present application is as follows: First, the industrial equipment original data stream is received, the feature vector is extracted through the protocol feature recognition model, and the dynamic protocol fingerprint library is matched; if it fails, the protocol learning mode is triggered: the response data stream is captured by actively sending a set of probe instructions, the frame structure features are analyzed to generate a protocol template, the analysis rules are derived in combination with the rule generator and the fingerprint library is updated. Then, the parsed structured data is input into the semantic understanding engine, the domain ontology library is queried based on the device ID and register address, and the physical quantity semantic label is labeled; according to the target system field description, the semantic similarity is calculated, and the conversion rule chain containing unit conversion and state code mapping is automatically generated; then, the CPU / memory / network parameters are monitored in real time, and when the CPU is greater than 80%, the lightweight mode is switched, and when the memory is less than 100MB, the aggregated calculation is offloaded to the cloud; finally, the data is encapsulated according to the target protocol and transmitted, and according to the feedback of the analysis timeout or field mismatch error code, the protocol frame header offset is corrected or the semantic similarity weight is adjusted.

[0022] In some optional embodiments, the protocol feature recognition model refers to an algorithm model for extracting key features of a protocol from a target device original data stream, such as frame structure, check bit, field offset, etc. It can be constructed based on machine learning or rule engine to realize automatic identification of protocol features.

[0023] For example, the CNN-based protocol feature recognition model extracts the feature vector from the original byte stream by analyzing the frame header 0x01, function code 0x03, data length, etc. of the Modbus protocol, which is used to match the known protocol fingerprint.

[0024] The protocol feature recognition model can be trained and generated through historical data streams of common protocols in the industrial field, and the model parameters can be iteratively optimized through feedback information of the dynamic protocol fingerprint library.

[0025] The dynamic protocol fingerprint library refers to a dynamically updated database that stores various industrial protocol feature templates, such as frame structure, check mode, register mapping rules, etc. It contains fingerprints and analysis rules of known protocols, and supports automatic storage of new protocols.

[0026] For example, the frame structure template of the Modbus protocol, the real-time frame format of the Profinet, and the custom frame features of a private protocol of a certain manufacturer are stored.

[0027] The dynamic protocol fingerprint library can be constructed based on pre-set industrial standard protocols; when a new protocol is added, the feature template and analysis rule are generated through the protocol learning mode and dynamically updated; based on the error code feedback of the target system (iterative optimization of the existing fingerprint.

[0028] Protocol learning mode refers to the operation mode of automatically learning unknown protocol features and generating parsing rules when dynamic protocol fingerprint library matching fails, without manual intervention to achieve new protocol adaptation.

[0029] For example, when encountering unknown private protocols, the system actively sends a set of probe instructions, captures device response data streams, analyzes their frame structure and check mode, and generates feature templates and register mapping rules for the protocol.

[0030] The acquisition method of the protocol learning mode is that when the feature vector extracted by the protocol feature recognition model has a matching degree lower than the threshold (such as 60%) with the dynamic protocol fingerprint library, the system is automatically triggered.

[0031] The domain ontology library refers to a structured knowledge base that stores the correlation of physical quantities, semantic tags, units, and register addresses in a specific industrial domain, and is used to realize semantic standardization of data.

[0032] For example, it contains semantic tags of physical quantities such as temperature and pressure, associated units, common register addresses such as 40001 corresponding to temperature sensors, and text descriptions of status codes such as running and fault shutdown.

[0033] The acquisition method of the domain ontology library is to pre-construct a basic ontology based on industry standards, supplement it with manual annotation of device manuals and expert knowledge, and dynamically adjust the correlation weight through semantic mapping failure.

[0034] The semantic understanding engine refers to a module that performs semantic annotation, unit conversion, and semantic mapping on structured source data based on the domain ontology library. Its core function is to eliminate semantic ambiguity of data and realize cross-system data fusion.

[0035] For example, the original value 255 of register address 40002 is annotated as pressure in combination with the domain ontology library, and the conversion logic chain is generated by matching the target system pipeline pressure field through cosine similarity calculation.

[0036] The acquisition method of the semantic understanding engine is to integrate natural language processing and semantic matching algorithms and rely on the domain ontology library. Its conversion logic chain can be corrected and optimized through manual annotation interface.

[0037] Lightweight, full-depth conversion mode refers to a conversion mode that only retains key data fields and compresses unnecessary information to reduce resource consumption when resources are scarce.

[0038] For example, when the CPU utilization rate exceeds 80%, only data with priority ≥3 such as temperature and fault alarms are processed, and secondary information such as device names and historical cumulative values is filtered, resulting in a data set compression rate of over 60%.

[0039] The acquisition mode of the lightweight full-depth conversion mode is triggered by the resource scheduling module according to real-time resource state parameters, and the lightweight full-depth conversion mode is automatically enabled when the CPU utilization rate is greater than or equal to a threshold value or the memory occupancy rate is greater than or equal to a threshold value.

[0040] The feedback optimization information refers to error codes, mapping failure records and other information returned by the target system, and is used for reverse optimization of the protocol fingerprint library and the domain ontology library to improve conversion accuracy.

[0041] For example, the target system returns an error of "0xE003 unit mismatch", indicating that there is a defect in the unit conversion rule in semantic mapping; and returns an error of "0xE001 parsing timeout", indicating that the frame structure template of the protocol fingerprint library is incorrect.

[0042] The acquisition mode of the feedback optimization information is that the protocol encapsulation and transmission module receives the feedback message of the target system, analyzes the error codes, log records and other information in the feedback message, and transmits the information to the protocol fingerprint learning module and the semantic understanding engine module for iterative optimization.

[0043] The raw data stream refers to the raw communication data collected directly from the communication interface of an industrial target device such as a sensor, a PLC or a smart instrument, without being subjected to protocol analysis or data conversion. The raw data stream exists in the form of binary byte stream, time sequence signal or electrical signal, and completely retains the underlying information such as frame structure, field order, check bit and idle timing of device communication.

[0044] For example, the raw hexadecimal byte stream sent by a temperature controller through RS485 in a serial port scenario is 01 03 0020 00 01 D5 CA, and the raw data stream in an Ethernet scenario and a wireless scenario; The acquisition mode of the raw data stream is to connect the device interface through an industrial communication adapter such as an RS485 to USB module or an Ethernet acquisition card, capture the electrical signal in real time and convert it into a binary byte stream.

[0045] The structured source data after analysis refers to the structured data obtained by applying the analysis rules of the dynamic protocol fingerprint library to the raw data stream, splitting the protocol frame structure, and converting the data segments into structured data with clear field definitions and data types. The structured source data after analysis is an intermediate product after protocol analysis.

[0046] The acquisition mode of the structured source data after analysis is to load the analysis rules of the dynamic protocol fingerprint library into the protocol analysis module, and perform frame splitting, field extraction and data type conversion on the raw data stream. If the raw data stream fails to match the fingerprint library, the protocol learning mode is triggered to generate analysis rules, and then the structured data is obtained by performing analysis.

[0047] Semantic data refers to the parsed structured source data, combined with the semantic association rules of the domain ontology library, giving the data physical quantity name, unit, state semantic label, business rules and other information, realizing the conversion of data from pure numerical / coding.

[0048] The acquisition method of semantic data is to load the domain ontology library, and perform the following on the structured source data: first, semantic query: find the physical quantity label according to the register address; then, unit conversion: convert the original numerical value to standard unit; finally, state code mapping: convert binary / hexadecimal state code to text description; if the semantic matching fails, record the feedback optimization information, and later through manual annotation or ontology library iteration to supplement the semantic rules.

[0049] Target protocol refers to the communication protocol supported by the target system, which specifies the data packaging format, transmission rules, and interaction logic, and is the data transmission standard after protocol conversion.

[0050] The acquisition method of target protocol is document query: refer to the technical manual and API document of the target system; protocol negotiation: determine the target protocol through the handshake process of the device and the target system; configuration extraction: extract the protocol format and field definition from the configuration file of the target system.

[0051] Embodiment two Please refer to Figures 1-6 The embodiment is further optimized on the basis of embodiment one. In this embodiment, the step of obtaining the matching result of the dynamic protocol fingerprint library according to the feature vector extracted by the protocol feature recognition model from the original data stream of the target device includes: According to the actively sent probe instruction set, obtain the interactive response data stream of the unknown protocol; According to the frame structure characteristics and check mode of the response data stream, generate a protocol feature template; According to the pre-set rule generator, derive the register mapping relationship, and obtain executable parsing rules.

[0052] In actual use, the step of obtaining the interactive response data stream of the unknown protocol according to the actively sent probe instruction set, sending a standard Modbus / OPC UA probe instruction set to the unknown device, capturing the response data stream and extracting the frame header positioning feature; the step of generating a protocol feature template according to the frame structure feature and the check mode of the response data stream, constructing a protocol feature template based on the fixed / variable length structure of the response data and the CRC check mode; the step of obtaining executable analysis rules by deriving the register mapping relationship according to the preset rule generator, analyzing the register address mapping rule by the preset rule generator, deriving the analysis rules with offset compensation, dynamically updating to the fingerprint library, and executing the method solves the problem that the traditional protocol analysis depends on the device manufacturer's documents, the manufacturer's document acquisition period is relatively long, and private protocols need to be reverse engineered, resulting in difficulty for small and medium-sized manufacturers to access.

[0053] In an aspect of the present disclosure, the step of obtaining semantic data with physical quantity labels by semantic annotation through a semantic understanding engine according to the analyzed structured source data and the domain ontology library comprises: querying the domain ontology library according to the device ID and the register address to obtain physical quantity semantic labels and units; automatically generating a conversion logic chain by calculating matching source data points according to the semantic description of the target system field through cosine similarity.

[0054] In actual use, the step of obtaining physical quantity semantic labels and units by querying the domain ontology library according to the device ID and the register address matches the register address to standard physical quantities and units according to the device ID associated with the domain ontology library; the step of automatically generating a conversion logic chain by calculating matching source data points according to the semantic description of the target system field through cosine similarity calculates the semantic similarity of the target field, and if it is greater than a threshold value, it is automatically bound to generate a conversion logic chain: calling the dimension formula library to convert Fahrenheit to Celsius, and converting 0x0001 to running through the state code mapping table. The advantage of this execution is that the automatic matching rate is higher, the unit conversion error rate is reduced, and the time consumption of cross-system field matching is shortened from hours to seconds; and after the method is executed, the problem of confusing units and state codes in manual configuration of the mapping table can be solved, and the situation of mapping failure due to naming differences between cross-system fields can be prevented.

[0055] In an aspect of the present disclosure, the step of automatically generating a conversion logic chain by calculating matching source data points according to the semantic description of the target system field through cosine similarity comprises: calling the dimension conversion formula library to obtain standard unit data according to the difference between the source data unit and the target unit; converting the binary state value to a text description readable by the target system according to the state code mapping table.

[0056] In actual use, when the source data unit (psi) and the target unit (MPa) are inconsistent, the formula library MPa=psi*0.00689 is called to perform real-time conversion; and the state code binary value, such as 0x0002, is converted into a text description by querying a preset mapping table, such as a fault shutdown. If an undefined code value is encountered, an artificial marking interface is triggered. After this method is executed, the problem of misjudgment of the control system caused by non-conversion and the problem of easy error of artificial maintenance conversion formula are effectively prevented, the precision of unit conversion is improved, the state code extension is real-time effective, and system failure caused by configuration error can be reduced.

[0057] In an aspect of the present disclosure, the step of obtaining semantic data with physical quantity labels by semantic annotation through a semantic understanding engine based on the parsed structured source data and the domain ontology library comprises: According to the result that the CPU utilization rate exceeds the threshold value, switching to a lightweight conversion mode to obtain simplified data that only retains key fields; According to the memory occupancy rate and the task calculation complexity evaluation, offloading the aggregation calculation task to the cloud to obtain a cross-node collaborative processing state code.

[0058] In actual use, when the CPU utilization rate exceeds the threshold value, the step of switching to a lightweight conversion mode to obtain simplified data that only retains key fields comprises: when the CPU is greater than 80%, starting the lightweight mode: only retaining data points with a priority greater than or equal to 3, and filtering 70% of low-importance data; and in the step of offloading the aggregation calculation task to the cloud to obtain a cross-node collaborative processing state code, evaluating the aggregation calculation complexity, such as 1.2 TFLOPS for Fourier transform, and if the memory is less than 100 MB, offloading the task to the cloud, and the edge node only forwarding instructions. The advantages of this execution are that, compared with the full conversion of the existing technology when the edge node resources are limited, which leads to data accumulation and fixed strategies cannot cope with load mutations, after the execution of the method in the embodiment, the resource consumption is reduced when the CPU is under high load, the task completion rate is significantly improved, and the cloud offloading increases the delay by less than 200 ms, but guarantees zero loss of key data.

[0059] In an aspect of the present disclosure, the step of switching to a lightweight conversion mode to obtain simplified data that only retains key fields based on the result that the CPU utilization rate exceeds the threshold value comprises: According to the data priority label, filtering low-importance data points to obtain a transmission data set with a compression rate of greater than or equal to 60%.

[0060] In an aspect of the present disclosure, the step of performing semantic labeling by a semantic understanding engine according to the parsed structured source data and the domain ontology library to obtain semantic data with physical quantity labels further comprises: According to the semantic label conflict result, the ontology library mapping relationship is corrected through a manual labeling interface to obtain a conflict-resolved conversion rule.

[0061] In actual use, when the semantic labels conflict, for example, 40001 in the ontology library is simultaneously mapped to temperature and voltage, the conflict point is popped up through the manual labeling interface, and a user correction instruction such as binding temperature is received, and the ontology library associated weight is updated: temperature weight + 0.7. This scheme solves the problem of high semantic conflict rate in the prior art when multiple source ontology libraries are merged, and the traditional scheme needs to be stopped for correction. Through the method in this embodiment, the conflict resolution of semantic labels is completed in real time, data loss is reduced, and the ontology library self-optimization greatly reduces the conflict probability.

[0062] In an aspect of the present disclosure, the step of performing semantic labeling by a semantic understanding engine according to the parsed structured source data and the domain ontology library to obtain semantic data with physical quantity labels further comprises: According to the network bandwidth fluctuation data, a Protobuf or JSON encapsulation format is dynamically selected to obtain a serialization message with an adaptive compression ratio.

[0063] As an optional implementation, in this embodiment, Protobuf serialization data is used, and the data volume is compressed by 65% through field label compression technology, and the number of key fields is ≤ 30% of the source data points, and the integrity of temperature, pressure and other core process parameters is ensured.

[0064] In an aspect of the present disclosure, the step of encapsulating data according to the semantic mapping result and the conversion strategy and transmitting the data through a priority queue to obtain feedback optimization information of a target system comprises: According to the target system error code feedback, the analysis rule defects of the protocol fingerprint library are located to obtain an iteratively optimized feature template; According to the semantic mapping failure record, the associated weight of the domain ontology library is adjusted to obtain an enhanced semantic matching accuracy.

[0065] In actual use, the step of locating the analysis rule defects of the protocol fingerprint library according to the target system error code feedback to obtain an iteratively optimized feature template is performed according to a 0xE001 analysis timeout error code, and the protocol frame header offset defect is located and corrected. The step of adjusting the associated weight of the domain ontology library according to the semantic mapping failure record to obtain an enhanced semantic matching accuracy is performed by analyzing the semantic mapping failure record and adjusting the associated weight of the ontology library.

[0066] In another aspect of the present disclosure, the present disclosure also relates to an industrial Internet of Things (IIoT) based data protocol conversion system, comprising a protocol fingerprint learning module, a semantic understanding engine module, a resource scheduling module, and a protocol encapsulation transmission module; the protocol fingerprint learning module is configured to obtain a dynamic protocol fingerprint library by extracting a feature vector from a raw data stream of a target device through a protocol feature recognition model and obtaining a matching result of the dynamic protocol fingerprint library; the semantic understanding engine module is configured to obtain semantic data with physical quantity labels by performing semantic labeling through a semantic understanding engine according to parsed structured source data and a domain ontology library; the resource scheduling module is configured to obtain an adaptive strategy and an offloading instruction by dynamically selecting a conversion strategy and deciding a calculation offloading point according to real-time resource state parameters; and the protocol encapsulation transmission module is configured to obtain a transmission result compatible with a target system.

[0067] In actual use, the feature extraction unit of the protocol awareness module extracts a protocol feature vector in real time, and the self-learning unit generates a new rule when a match fails. The ontology labeling unit of the semantic engine module binds a physical quantity label, and the rule generation unit outputs a mapping logic. The resource monitoring unit of the dynamic scheduling module reports resource data every second, the strategy selector switches a conversion mode, and the offloading controller allocates a cloud task. The encapsulation unit of the execution optimization module serializes data, and the error analysis unit drives double-library iteration.

[0068] As an optional implementation, the present embodiment is only used as an example of an optional implementation; Embodiment Three As shown in Figures 1-6 The present embodiment is further optimized on the basis of Embodiment One and Embodiment Two. After the step of encapsulating data according to the semantic mapping result and the conversion strategy, and transmitting the data through a priority queue to obtain feedback optimization information of a target system is performed, the following steps are further performed: generating a protocol defect topology graph according to an error code type of the target system; The protocol analysis error is associated with the step of obtaining a matching result of a dynamic protocol fingerprint library by extracting a feature vector from a raw data stream of a target device through a protocol feature recognition model, and the protocol fingerprint library feature template of the protocol awareness module; The semantic mapping error is associated with the weight of the domain ontology library in the step of obtaining semantic data with physical quantity labels by performing semantic labeling through a semantic understanding engine according to parsed structured source data and a domain ontology library.

[0069] The matching priority of the protocol analysis core is dynamically adjusted based on the historical error occurrence frequency, and the optimization formula is configured as:

[0070] wherein, and is an adjustable coefficient, = 0.3; is the protocol analysis core matching priority; is the historical analysis error number; is the semantic association weight.

[0071] In some different embodiments, the step of dynamically selecting a conversion strategy and deciding a calculation offloading point to obtain lightweight and full-depth conversion instructions according to real-time resource state parameters includes: An edge cloud collaborative scoring model is constructed, and the model is configured to:

[0072] wherein, is the edge cloud collaborative score; is the available bandwidth; is the vibration spectrum task complexity; When is less than 0, the FFT analysis task is offloaded to the cloud K8s cluster.

[0073] In some different embodiments, the step of obtaining semantic data with physical quantity labels by semantic annotation through a semantic understanding engine according to the parsed structured source data and the domain ontology library further includes: When an undefined register address is detected, cross-protocol field similarity calculation is started; Temporary semantic mapping rules are generated and pushed to an engineer terminal for confirmation; After confirmation, the domain ontology library association weight is dynamically updated.

[0074] As an optional implementation, for example, when the wind power monitoring platform frequently returns "0xE003 - unit mismatch" error; First, protocol defect topology optimization is performed, in which process the error analysis module locates semantic layer defects, then the ontology library is traced to find that the "temperature" field is simultaneously bound to ℃ / ℉; then the error frequency is dynamically adjusted: the ℉ unit error rate is 18%, and the weight is reduced to 0.2; the ℃ unit error rate is 0.5%, and the weight is increased to 0.8; the subsequent conversion error rate is reduced from 15% to 0.7%.

[0075] Then edge cloud collaborative computing is performed Detect edge node status: CPU utilization = 92%, available bandwidth = 8Mbps, vibration spectrum task complexity = 12.5TFLOPS, finally calculate through edge cloud collaborative scoring model:

[0076] Since Less than 0, so execute: Local only keep key fields; FFT analysis task offloading to AWS c6g instance; Finally, the edge node load is from 92% to 58%, and the task latency increases by 182ms (<ISO 22400 standard 200ms threshold).

[0077] It should be further explained that the whole data protocol conversion system based on industrial Internet of Things can be applied to the optimized industrial Internet of Things, such as Figure 5 As shown, the optimized industrial Internet of Things includes a user platform, a service platform, a management platform, a sensing network platform and an object platform which are sequentially established in communication; The user platform is configured to provide the function of front-end service to the user; the user obtains the required sensing service information through the user platform, processes the sensing service information, and converts it into user sensing information; the user analyzes the user sensing information and makes corresponding decisions combined with his own will, converts the user sensing information into user control information through the corresponding information system and sends it to the service platform, thereby expressing the corresponding service demand will of the user.

[0078] The physical entity of the user platform includes various user terminals such as mobile phones, computers, special terminals, etc., which realize the service of the user terminal through the combination with the user information system software.

[0079] The service platform is configured as an API server or other server for establishing communication between the management platform and the user platform to realize the corresponding function; the physical entity of the service platform includes various servers.

[0080] The management platform is configured to perform at least one of device running state monitoring management, data monitoring management, device parameter management, and life cycle management; the management platform is an operation and coordination platform of the Internet of Things, which can include various management sub-platforms, and different management sub-platforms perform different management businesses; the physical entity of the management platform includes various servers.

[0081] The sensing network platform is configured to perform at least one of network management, instruction management, device state management, data protocol management, data analysis, data classification, data transmission monitoring, and data transmission security management. The sensing network platform provides functions such as communication transmission, analysis, identification, and classification of data, avoiding direct aggregation of various object platform data on the management platform, causing the management platform data to be redundant, and the data processing efficiency to be low. The physical entity of the object platform includes various gateways, edge computing devices, and the like.

[0082] The object platform is configured to perform specific production control, detection, measurement, and other production work. The physical entity of the production object includes various production devices, sensors, and the like.

[0083] Figure 6 is a block diagram of an electronic device according to an example embodiment of an industrial Internet of Things-based data protocol conversion method. As shown in Figure 6 the electronic device 700 can include a processor 701, a memory 702. The electronic device 700 can also include one or more of a multimedia component 703, an I / O interface 704 (input / output interface), and a communication component 705.

[0084] The processor 701 is configured to control overall operations of the electronic device 700 to complete all or part of the steps of the above-described data protocol conversion method based on industrial Internet of Things. The memory 702 is configured to store various types of data to support operations of the electronic device 700, which can 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, pictures, audio, video, and the like. 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, magnetic disk or optical disk. The multimedia component 703 can include a screen and an audio component. The screen can be, for example, a touch screen, and the audio component is configured to output and / or input audio signals. For example, the audio component can include a microphone configured to receive external audio signals. The received audio signals can be further stored in the memory 702 or transmitted through the communication component 705. The audio component also includes at least one speaker configured to output audio signals. The I / O interface 704 provides an interface between the processor 701 and other interface modules, which can be a keyboard, a mouse, a button, and the like. The buttons can be virtual buttons or physical buttons. The communication component 705 is configured to perform 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, and the like, or a combination of one or more of them, is not limited herein. Therefore, the corresponding communication component 705 can include a Wi-Fi module, a Bluetooth module, an NFC module, and the like.

[0085] In an exemplary embodiment, the electronic device 700 can be implemented by one or more Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic elements for performing the above-described industrial internet of things based data protocol conversion method.

[0086] In another exemplary embodiment, a computer readable storage medium including program instructions which, when executed by a processor, implement the steps of the above-described industrial internet of things based data protocol conversion method is also provided. For example, the computer readable storage medium can be the above-described memory 702 including program instructions which can be executed by the processor 701 of the electronic device 700 to complete the above-described industrial internet of things based data protocol conversion method.

[0087] In another exemplary embodiment, a computer program product containing a computer program which is executable by a programmable apparatus and which has code portions for performing the above-described industrial internet of things based data protocol conversion method when executed by the programmable apparatus is also provided.

[0088] In the description of the present application, it should be understood that the terms "coaxial", "bottom", "one end", "top", "middle", "the other end", "upper", "one side", "top", "inner", "front", "central", "both ends" and the like indicate the orientation or positional relationship based on the orientation or positional relationship shown in the drawings, and are only for the purpose of facilitating the description of the present application and simplifying the description, and do not indicate or imply that the indicated device or element must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present application.

[0089] In addition, the terms "first", "second", "third", "fourth" are only for the purpose of description, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features, so that the features with "first", "second", "third", "fourth" can explicitly or implicitly include at least one of the features.

[0090] In the present application, unless otherwise explicitly specified and limited, the terms "mounting", "setting", "connecting", "fixing", "screwing" and other terms should be understood in a broad sense, for example, can be fixed connection, can also be detachable connection, or integrated; can be mechanical connection, can also be electrical connection; can be directly connected, can also be indirectly connected through an intermediate medium, can be the internal communication of two elements or the interaction relationship of two elements, unless otherwise explicitly limited, the person skilled in the art can understand the specific meaning of the above terms in the present application according to the specific circumstances.

[0091] Although the embodiments of the present application have been shown and described, it can be understood by those skilled in the art that various changes, modifications, replacements and variations can be made to the embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A data protocol conversion method based on the Industrial Internet of Things, characterized in that, include: Based on the raw data stream of the target device, feature vectors are extracted through the protocol feature recognition model to obtain the matching results of the dynamic protocol fingerprint database; Based on the matching results of the dynamic protocol fingerprint database, determine whether to trigger the protocol learning mode to obtain new fingerprints and parsing rules; Based on the structured source data parsed according to the parsing rules and the domain ontology library, semantic annotation is performed through the semantic understanding engine to obtain semantic data with physical quantity labels; Based on real-time resource status parameters, the conversion strategy is dynamically selected and the unloading point is calculated to obtain lightweight, full-scale, and deep conversion instructions. Based on the semantic mapping results and the conversion strategy, data is encapsulated according to the target protocol and transmitted through a priority queue to obtain feedback optimization information from the target system.

2. The data protocol conversion method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of extracting feature vectors from the target device's original data stream using a protocol feature recognition model to obtain the matching results of the dynamic protocol fingerprint database includes: Based on the actively sent set of probe commands, obtain the interactive response data stream of the unknown protocol; Based on the frame structure characteristics and verification mode of the response data stream, a protocol feature template is generated; The register mapping relationship is derived from the preset rule generator to obtain executable parsing rules.

3. The data protocol conversion method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of obtaining semantically labeled data with physical quantity tags by performing semantic annotation through a semantic understanding engine based on the parsed structured source data and domain ontology library includes: Query the domain ontology library based on the device ID and register address to obtain the semantic tags and units of physical quantities; Based on the semantic description of the target system fields, the source data points are matched using cosine similarity calculation to obtain an automatically generated transformation logic chain.

4. The data protocol conversion method based on the Industrial Internet of Things according to claim 3, characterized in that: The step of obtaining an automatically generated transformation logic chain by calculating the matching source data points through cosine similarity based on the semantic description of the target system fields includes: Based on the difference between the source data units and the target units, the unit conversion formula library is called to obtain standard unit data; Based on the status code mapping table, the binary status value is converted into a text description readable by the target system.

5. The data protocol conversion method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of obtaining semantically labeled data with physical quantity tags by performing semantic annotation through a semantic understanding engine based on the parsed structured source data and domain ontology library includes: Based on the result that the CPU utilization exceeds the threshold, switch to lightweight transformation mode to obtain simplified data that retains only the key fields; Based on the assessment of memory usage and task computational complexity, the aggregated computing task is offloaded to the cloud to obtain the cross-node collaborative processing status code.

6. The data protocol conversion method based on the Industrial Internet of Things according to claim 5, characterized in that: The step of switching to lightweight conversion mode based on the result that CPU utilization exceeds a threshold to obtain simplified data that retains only key fields includes: Based on data priority labels, low-importance data points are filtered out to obtain a transmission dataset with a compression rate of ≥60%.

7. The data protocol conversion method based on the Industrial Internet of Things according to claim 2, characterized in that: The step of obtaining semantically labeled data with physical quantity tags by performing semantic annotation through a semantic understanding engine based on the parsed structured source data and domain ontology library further includes: Based on the semantic tag conflict results, the ontology mapping relationship is corrected through a manual annotation interface to obtain the conversion rules after conflict resolution.

8. The data protocol conversion method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of obtaining semantically labeled data with physical quantity tags by performing semantic annotation through a semantic understanding engine based on the parsed structured source data and domain ontology library further includes: Based on network bandwidth fluctuation data, either Protobuf or JSON can be dynamically selected as the encapsulation format to obtain a serialized message with an adaptive compression ratio.

9. The data protocol conversion method based on the Industrial Internet of Things according to claim 1, characterized in that: The step of encapsulating data according to the target protocol and transmitting it through a priority queue based on the semantic mapping result and the conversion strategy to obtain feedback optimization information of the target system includes: Based on the error code feedback from the target system, locate the defects in the parsing rules of the protocol fingerprint database and obtain the iteratively optimized feature template; Based on the semantic mapping failure records, the association weights of the domain ontology are adjusted to obtain enhanced semantic matching accuracy.

10. A data protocol conversion system based on the Industrial Internet of Things, characterized in that, include: The protocol fingerprint learning module is configured to obtain a dynamically updated protocol fingerprint database by extracting feature vectors from the target device's original data stream using a protocol feature recognition model. The semantic understanding engine module is configured to perform semantic annotation based on the parsed structured source data and the domain ontology library, thereby obtaining semantic data with physical quantity labels, and obtaining semantic data and conversion rules. The resource scheduling module is configured to obtain the adaptive strategy and unloading instruction by dynamically selecting the conversion strategy and deciding the calculation of the unloading point based on the real-time resource status parameters and obtaining lightweight and full-depth conversion instructions. as well as The protocol encapsulation and transmission module is configured to obtain a transmission result compatible with the target system by dynamically selecting a conversion strategy and deciding on the offload point based on real-time resource status parameters to obtain lightweight, full-scale deep conversion instructions.

Images