Industrial computer multi-protocol adaptive control system based on edge computing

Abstract

An industrial computer multi-protocol adaptive control system based on edge computing relates to the technical field of industrial automation control, and performs three-level progressive protocol identification on new industrial equipment accessed to an industrial network diagram to obtain a communication protocol of the new industrial equipment or mark the new industrial equipment as a protocol unknown state; constructing a reverse representation tree of each communication protocol, performing protocol analysis on the new industrial equipment marked as a protocol unknown state based on the reverse representation tree, and performing protocol analysis on the new industrial equipment with a known communication protocol based on a Docker container mirror image; performing dynamic protocol conversion on the analysis result of the new industrial equipment according to the real-time state data of the new industrial equipment and a preset protocol conversion rule; setting a self-adaptive buffer area and a dynamic priority of the target equipment according to a communication protocol of the new industrial equipment; temporary dynamic feedback detection is carried out on the new industrial equipment provided with the temporary dynamic feedback detection mechanism, and the accuracy and efficiency of the multi-protocol identification process are remarkably improved.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial automation control, and specifically to a multi-protocol adaptive control system for industrial computers based on edge computing. Background Art

[0002] A Chinese patent with publication number CN119232763A discloses an industrial control loop network monitoring system and its monitoring method, including: a data acquisition module: obtaining industrial control loop network data packets in the industrial control network; a data processing module: performing data cleaning and data sorting on the collected data, and performing protocol matching on the data packets to identify the industrial protocols in the data packets; a data identification module: performing protocol matching on the data in the network data packets to identify the communication data of each network packet. A Chinese patent with publication number CN115190191A discloses a power grid industrial control system and a control method based on protocol parsing, including: obtaining in real time the protocol data stream generated by the communication between the controller and the upper computer of the power grid industrial control system; performing periodic parsing on the obtained protocol data stream, and encapsulating and reporting; performing de-encapsulation on the parsed result of the encapsulated and reported data, monitoring the behavior characteristics of the specified fields of the de-encapsulated protocol data stream, and determining whether the protocol data stream is abnormal; when it is determined that the protocol data stream is abnormal, terminating the control process of the power grid industrial control system, alarming and recording logs, and transmitting the alarm information and log information to the terminal module.

[0003] Traditional industrial control systems rely on predefined protocol libraries, making it difficult to dynamically identify and be compatible with unknown protocol devices. For new devices, protocol parameters need to be manually configured, resulting in low efficiency and poor scalability. Traditional systems lack an adaptive mechanism and cannot dynamically adjust buffer capacity and priority according to device communication characteristics, easily leading to data congestion or key service delays. When multiple devices communicate concurrently, network performance is unstable, affecting system reliability. At the same time, the existing systems' parsing of unknown protocols relies on manual reverse engineering, which is time-consuming and costly, and it is difficult to quickly respond to the access requirements of new devices. Enterprises need to develop parsing modules separately for each new protocol, extending the system deployment cycle. At the same time, the existing systems lack real-time monitoring and dynamic feedback of device communication status, cannot detect protocol misjudgments or network anomalies in a timely manner, and the protocol conversion rules are fixed and cannot dynamically adjust the conversion strategy according to the real-time status of the device. Summary of the Invention

[0004] To solve the above technical problems, the purpose of the present invention is to provide an industrial computer multi-protocol adaptive control system based on edge computing, including an industrial computer control center and edge computing nodes. The industrial computer control center is distributively connected to a number of industrial devices through the edge computing nodes. The edge computing nodes include a protocol recognition module, a protocol parsing module, a protocol conversion module, an adaptive buffer management module, and a dynamic feedback detection module; The protocol recognition module is used to perform CRC matching and address feature Bayesian classification on newly added industrial devices accessing the industrial network diagram, and set a temporary dynamic feedback detection mechanism or perform lightweight feature matching on the newly added industrial devices according to the matching results, so as to obtain the communication protocols of the newly added industrial devices or mark the newly added industrial devices as in an unknown protocol state; The protocol parsing module is used to construct an inverse representation tree for each communication protocol, obtain the tree structure features, node attribute features, sequence features, and Docker container images of each communication protocol, and perform protocol parsing on the newly added industrial devices marked as in an unknown protocol state based on the inverse representation tree, and perform protocol parsing on the newly added industrial devices with known communication protocols based on the Docker container images; The protocol conversion module is used to collect the real-time status data of the newly added industrial devices, and perform dynamic protocol conversion on the parsing results of the newly added industrial devices according to the real-time status data and the preset protocol conversion rules; The adaptive buffer management module is used to set an adaptive buffer area and dynamic priority for the target device according to the communication protocol of the newly added industrial devices; The dynamic feedback detection module is used to perform temporary dynamic feedback detection on the newly added industrial devices with a temporary dynamic feedback detection mechanism set.

[0005] Furthermore, obtain the communication connection relationships between a number of industrial devices. Take the number of industrial devices as nodes and the communication connection relationships between the number of industrial devices as the connection relationships between the nodes to construct an industrial network diagram. Perform real-time detection of newly added industrial device access on the industrial network diagram. Use the network scanning tool Nmap to scan the IP address segments of each industrial device in the industrial network diagram to identify the IP addresses of the existing industrial devices and the IP addresses of the newly added industrial devices. When a newly added industrial device is detected to be accessing the industrial network diagram, pack the data generated by the newly added industrial device into data packets and upload them to the protocol recognition module.

[0006] Furthermore, the process of the protocol recognition module performing CRC matching and address feature Bayesian classification on the newly added industrial devices accessing the industrial network diagram and setting a temporary dynamic feedback detection mechanism or performing lightweight feature matching on the newly added industrial devices according to the matching results includes: Construct an open protocol fingerprint database, where the open protocol fingerprint database includes CRC features and address features corresponding to several open protocols. The CRC features include CRC generation polynomials, initial values, check bit lengths, etc., and the address features include address values, address ranges, encoding rules, etc.; Pack the data generated by the new industrial device into data packets, extract features from the data packets to obtain target CRC features and target address features, input the target CRC features into the open protocol fingerprint database for CRC matching, screen out the open protocols in the open protocol fingerprint database whose corresponding CRC features are consistent with the target CRC features, and perform Bayesian classification of address features on the open protocols according to the target address features to obtain the posterior probabilities of the target address features belonging to each open protocol; Among them, the specific process of performing Bayesian classification of address features on the open protocols according to the target address features to obtain the probabilities of the target address features belonging to each open protocol includes: Let C be the open protocol category and X be the target address feature (such as address value, address range, encoding rule, etc.). Convert the target address feature X into a discrete feature vector. For example, the address range (4x / 3x / 0x / 1x), register type (holding register / input register), slave address (0x0100 - 0x01FF), PDO mapping address (0x1C10 - 0x1C13), and vectorize the target address feature into , where is a binary feature (such as "whether it belongs to the Modbus4x register"), and n represents the number of discrete feature vectors; Since the address features of each open protocol are independent of each other, construct a simplified posterior probability formula: ; Among them, represents the posterior probability that the target address feature X belongs to the open protocol C, is the prior probability, representing the appearance probability of the open protocol C in the industrial network diagram, is the likelihood, representing the probability that the open protocol C generates the target address feature ; Preset a posterior probability determination threshold, screen out the open protocol with the largest posterior probability, compare the posterior probability of the open protocol with the posterior probability determination threshold. If the posterior probability of the open protocol is greater than or equal to the posterior probability determination threshold, label the open protocol as the communication protocol of the new industrial device and set up a temporary dynamic feedback detection mechanism for the new industrial device. If the posterior probability of the open protocol is less than the posterior probability determination threshold, perform lightweight feature matching; If there is no open protocol in the open protocol fingerprint database whose corresponding CRC features are consistent with the target CRC features, perform lightweight feature matching.

[0007] Furthermore, the process of lightweight feature matching includes: Collect a number of consecutive data packets generated by the new industrial device, perform multi-dimensional feature extraction on the number of consecutive data packets to generate a time feature vector, a structure feature vector, and a data content feature vector, and perform feature extraction according to the CRC matching result and the address feature Bayesian classification result of the new industrial device to obtain protocol fingerprint features; Build a private protocol recognition model based on lightweight CNN+LSTM, input the time feature vector, the structure feature vector, the data content feature vector, and the protocol fingerprint features into the private protocol recognition model, and obtain the confidence levels of a number of private protocols according to the output of the private protocol recognition model; Preset a confidence threshold, screen out the private protocol with the highest confidence level, compare the confidence level of the private protocol with the confidence threshold. If the confidence level is greater than or equal to the confidence threshold, mark the private protocol as the communication protocol of the new industrial device. If the confidence level is less than the confidence threshold, mark the new industrial device as in a protocol unknown state.

[0008] Among them, the time feature vector includes the mean frame interval, the standard deviation of the frame interval, the maximum frame interval, and the minimum frame interval; the structure feature vector includes the frame length and the number of fields; the data content feature vector includes the data field entropy value, the occurrence frequency of special values (such as 0xFF (Modbus broadcast address), 0x0000 (initial value)); the protocol fingerprint features include the matching degree between the target CRC feature and the CRC features of each public protocol in the public protocol fingerprint library, and the posterior probability of each public protocol; Furthermore, in the process of performing feature extraction according to the CRC matching result of the new industrial device, a hierarchical scoring method is used to obtain the matching degree between the target CRC feature and the CRC features of each public protocol in the public protocol fingerprint library. For example: assign weights according to the importance of parameters (example weights: polynomial 40%, number of bits 20%, byte order 20%, initial value 15%, check position 5%) to obtain the target CRC feature. Step 1: Match the number of bits. The number of bits must be the same, otherwise it is directly excluded (such as if the target is a 16-bit CRC and the CAN bus (15 bits) has a matching degree of 0%). Step 2: Match the polynomial. If the polynomials are exactly the same, full marks are obtained. Otherwise, points are deducted according to the Hamming distance (the number of different bits in binary). (If the target polynomial 0x8005 (binary 1000 0000 0000 0101) is exactly the same as Modbus, the score is 100%; it has 6 differences from 0x1021 (0001 0000 0010 0001) of Zigbee, and the score is 1 - 6 / 16 = 62.5%). Step 3: Match the byte order, initial value, and check position. Byte order: If the high - low bit order is the same, the score is 100%, otherwise 0%. Initial value: If the hexadecimal values are the same, the score is 100%, otherwise, points are deducted according to the difference ratio. Check position: If the positions are the same (e.g., both are at the end of the frame), the score is 100%, otherwise 0%. Step 4: Calculate the total matching degree by weighted calculation, 。

[0009] Further, the process of the protocol parsing module constructing the reverse representation trees of each communication protocol and obtaining the tree - structure features, node - attribute features, sequence features of each communication protocol, and Docker container images includes: Construct a protocol parsing database, obtain the parsing logics of several communication protocols (including public protocols and private protocols). The parsing logics include parsing fixed fields (such as slave address, function code) and variable fields (such as register address, data value) of data packets, etc. Package the parsing logic of each communication protocol into a Docker container image, and the Docker container image contains independent code, dependency libraries, and configuration files; Obtain sample data packets corresponding to several communication protocols; Perform field segmentation and hierarchical analysis on the sample data packets of the communication protocols, obtain the nature of each field in the sample data packets and the hierarchical relationship between each field. Define different types of nodes according to the nature and hierarchical relationship of the fields, create corresponding instance nodes for each field and set the attributes of the instance nodes (basic field → basic node, set attributes such as type, length, etc. Composite field → composite node, set the list of child nodes. Check field → check node, specify the check algorithm and associated fields). According to the hierarchical relationship between each field, determine the hierarchical relationship between the corresponding instance nodes created for each field. Starting from the root node, connect each instance node according to the hierarchical relationship to generate the reverse representation tree of the communication protocol. Extract features from the reverse representation tree of the communication protocol to generate tree - structure features, node - attribute features, and sequence features; Associate and store the tree - structure features, node - attribute features, and sequence features of each communication protocol with the Docker container images of each communication protocol in the protocol parsing database.

[0010] Further, the process of protocol parsing for new industrial devices marked as in an unknown protocol state based on the reverse representation tree and protocol parsing for new industrial devices with known communication protocols based on Docker container images includes: If the new industrial device is marked as having an unknown protocol status, the data packets of the new industrial device are segmented by fields and analyzed hierarchically to construct an inverse representation tree corresponding to the data packets of the new industrial device. Feature extraction is performed on the inverse representation tree to generate tree structure features, node attribute features, and sequence features. The tree structure features, node attribute features, and sequence features are input into the protocol parsing database for matching to obtain the representation tree similarity of each communication protocol in the protocol parsing database. The communication protocol with the highest representation tree similarity is selected and marked as the communication protocol of the new industrial device. The Docker container image of the communication protocol is downloaded from the protocol parsing database, and the data packets of the new industrial device are parsed according to the Docker container image to obtain the key data in the data packets. For example, for the Modbus protocol, parsing the data packets according to the Docker container image can accurately extract key data such as register addresses and data values from the data packets, providing a basis for subsequent data processing and control; The calculation formula for inputting the tree structure features, node attribute features, and sequence features into the protocol parsing database for matching to obtain the representation tree similarity of each communication protocol in the protocol parsing database is as follows: Preset the tree structure feature T, , where d represents the depth of the tree, n represents the total number of instance nodes, b represents the average branching factor (number of leaf nodes / number of non-leaf nodes), represents the number of instance nodes in the i-th layer, represents the total number of layers; Preset the attribute feature of each node as A, , where t represents the field type (such as Int, String, encoded as a vector through one-hot), l represents the field length (such as the number of bytes), r represents the value range ([0, 255] in this example, normalized to 0 - 1), represents whether it is a reserved field (0 or 1), which can be merged into the encoding of t (for example, adding a Reserved type); Preset the field type sequence as S, , where m represents the number of field types; ; ; ; ; Among them, represents the structure similarity of two inverse representation trees and , represents the feature dimension, such as =5 corresponding to , Represents the most average attribute similarity corresponding to all instance nodes traversing two reverse representation trees. Represents the number of instance nodes where two reverse representation trees match. Represents the type weight. Represents the vector cosine similarity. Used to measure the direction consistency of two attribute vectors (mainly t and r). Represents the normalized absolute difference, used to measure the length difference, and the result range is [0,1]. Represents the sequence feature similarity. Represents the field type sequence and the length of the longest common subsequence; Represents the representation tree similarity between two reverse representation trees. 、 and are weight factors; If the new industrial device is not marked as the protocol unknown state, obtain the communication protocol of the new industrial device, download the Docker container image of the communication protocol of the new industrial device from the protocol parsing database, and parse the data packets of the new industrial device according to the Docker container image to obtain the key data in the data packets.

[0011] When the new industrial device cannot be detected in the industrial network diagram, delete the Docker container image of the communication protocol corresponding to the new industrial device.

[0012] Furthermore, the protocol conversion module collects the real-time status data of the new industrial device. The process of dynamically converting the parsing result of the new industrial device according to the real-time status data and the preset protocol conversion rules includes: Construct an intermediate conversion model, obtain the industrial device communicating with the new industrial device, mark the industrial device as the target device, pre-construct a device information database, and the device information database is used to record the communication protocol types supported by each industrial device, and obtain the communication protocol type supported by the target device according to the device information database; Preset several protocol conversion rules, and each protocol conversion rule includes a conversion condition and a target protocol. The conversion conditions include network load, device type, and device status. Collect the real-time status data of the new industrial device and match the real-time status data of the new industrial device with the conversion conditions of each protocol conversion rule; If there is a protocol conversion rule whose corresponding conversion condition matches the real-time status data. For example, Rule 1: If the network load < 30% and the device is a PLC, then execute Modbus→MQTT (high bandwidth efficiency); Rule 2: If the device power < 20%, execute Modbus→CoAP (low power consumption). Obtain the target protocol of the new industrial device according to the protocol conversion rule, and determine whether the communication protocol type supported by the target device includes the target protocol of the new industrial device. If it does not include, perform a two-way protocol conversion operation. If it includes, input the target protocol and the key data in the data packet into the intermediate conversion model, and according to the intermediate conversion model, output the target protocol representation form of the key data, and transmit the target protocol representation form of the key data to the target device; If there is no protocol conversion rule whose corresponding conversion condition matches the real-time status data, then determine whether the communication protocol types of the new industrial device and the target device are the same. If they are the same, transmit the key data to the target device. If they are not the same, input the key data in the data packet and the communication protocol of the target device into the intermediate conversion model, and according to the intermediate conversion model, output the communication protocol representation form of the key data and transmit it to the target device.

[0013] Further, the process of performing the two-way protocol conversion operation includes: Input the target protocol of the new industrial device and the key data in the data packet into the intermediate conversion model, and according to the intermediate conversion model, output the target protocol representation form of the key data. Upload the target protocol representation form of the key data to the industrial computer control center, and according to the industrial computer control center, convert the target protocol representation form of the key data into the communication protocol representation form of the target device, and then transmit the communication protocol representation form of the key data to the target device.

[0014] Further, the process of the adaptive buffer management module setting the adaptive buffer area and dynamic priority of the target device according to the communication protocol of the new industrial device includes: Pre-set the initial capacity of the independent buffer for different communication protocols. If the communication protocol of the target device is inconsistent with the communication protocol of the new industrial device, set an independent buffer for the target device according to the initial capacity of the independent buffer of the communication protocol of the new industrial device, monitor the protocol metrics of the independent buffer of the target device, and obtain the protocol metrics. The protocol metrics include: latency: the time from when the data enters the buffer to when it exits the buffer; queue length: the number of data packets to be processed in the current buffer. Preset the protocol metric threshold interval, and compare the protocol metrics with the protocol metric threshold interval. If the protocol metrics are not within the protocol metric threshold interval, perform an adaptive capacity adjustment on the independent buffer; Among them, the specific formula for performing the adaptive capacity adjustment on the independent buffer is: ; Among them, represents the new capacity, represents the current capacity, represents the delay threshold, represents the delay, represents the queue length, represents the queue length threshold, , are the weight coefficients of the delay and the queue length respectively, and ; If the target device is simultaneously communicatively connected to k (k > 1) new industrial devices, and the communication protocols of the target device are all inconsistent with those of the k new industrial devices, then k independent buffers are set for the target device according to the initial capacity of the independent buffer of each new industrial device's communication protocol. The protocol metrics of each independent buffer and the protocol real-time level of the communication protocol corresponding to each independent buffer are obtained, and the dynamic priority of each independent buffer is obtained according to the protocol metrics and the protocol real-time level of each independent buffer.

[0015] Among them, the specific formula for obtaining the dynamic priority of each independent buffer is: ; Among them, represents the dynamic priority, represents the protocol real-time level.

[0016] Furthermore, the process of the dynamic feedback detection module performing temporary dynamic feedback detection on the new industrial device with the temporary dynamic feedback detection mechanism set includes: Setting a feedback detection period for the new industrial device with the temporary dynamic feedback detection mechanism set. During the feedback detection period, the data logs of the new industrial device and the target device are collected, and the data logs are statistically analyzed to obtain the interaction metrics between the new industrial device and the target device. The interaction metrics include data loss frequency, data delay, connection establishment time, disconnection and reconnection times, function execution results, etc.; Presetting an interaction metric threshold interval. If the interaction metrics of the new industrial device are not within the interaction metric threshold interval, then the communication protocol of the new industrial device is used to perform lightweight feature matching on the new industrial device; If the interaction metrics of the new industrial device are within the interaction metric threshold interval, then the temporary dynamic feedback detection mechanism of the new industrial device is removed.

[0017] Compared with the prior art, the beneficial effects of the present invention are: 1. Efficient protocol recognition and adaptive detection mechanism CRC Matching and Bayesian Classification: By matching the CRC features and address features in the public protocol fingerprint library and combining with Bayesian posterior probability determination, the recognition accuracy of known protocols is increased to over 98.7% (85% - 90% for traditional methods). For example, the recognition time for the Modbus protocol is shortened from the traditional 300ms to 80ms.

[0018] Lightweight Feature Matching: For unknown protocol devices, through multi-dimensional analysis of time features, structural features, and data content features, combined with a private protocol recognition model, the parsing success rate of unknown protocols is increased to 92% (75% for traditional rule-based methods). For example, the recognition speed for a certain private industrial protocol is shortened from 2 hours to 15 minutes.

[0019] Dynamic Feedback Detection: Through a temporary dynamic feedback detection mechanism, the communication stability of unknown protocol devices is verified in real time, reducing system failures caused by misjudgments. For example, the communication anomaly detection period for a certain sensor device is shortened from 4 hours to 15 minutes.

[0020] 2. Intelligence and Standardization of Protocol Parsing Reverse Representation Tree and Docker Containerization: Encapsulate the protocol parsing logic into a Docker image to achieve plug-and-play of the parsing module and reduce the system upgrade cost. For example, the deployment time for adding the OPC UA protocol parsing function is shortened from the traditional 2 weeks to 2 hours.

[0021] Tree Structure Feature Matching: Through matching the structural features, node attribute features, and sequence features of the protocol reverse representation tree, rapid parsing of unknown protocols is achieved. For example, the parsing time for the BACnet protocol is shortened from the traditional 400ms to 120ms.

[0022] Intermediate Conversion Model: By standardizing the intermediate data model, the complexity of protocol conversion is reduced by over 60%. For example, the code volume for two-way conversion between Modbus and MQTT is reduced from the traditional 2000 lines to 500 lines.

[0023] 3. Real-time Performance and Compatibility of Dynamic Protocol Conversion Two-way Protocol Conversion: Support a hybrid mode of local conversion at the edge node and collaborative conversion in the cloud to reduce data transmission latency. For example, local conversion at the edge node can achieve a latency of 1ms, and the average latency of cloud conversion is reduced from 500ms to 150ms.

[0024] Real-time State Awareness: Dynamically select the optimal conversion path based on the real-time state of the device (such as network load, device power) to improve system resource utilization. For example, when the network load is higher than 80%, automatically select the Protobuf protocol with a higher compression rate to reduce the data transmission volume by 40%.

[0025] Multi - protocol adaptation: It supports the conversion of more than 10 heterogeneous protocols simultaneously. For example, in a smart factory scenario, it realizes the mixed communication of protocols such as Modbus, EtherCAT, and MQTT, and the system throughput is increased by 300%.

[0026] 4. Reliability and efficiency of adaptive buffer management Independent buffer and dynamic priority: Allocate independent buffers for each protocol, and combine dynamic priority scheduling to control the latency of real - time protocols (such as EtherCAT) within 2ms, and reduce the packet loss rate from the traditional 12% to 0.05%.

[0027] Dynamic capacity adjustment: Based on the adaptive adjustment algorithm of latency and queue length, keep the buffer utilization rate always in the optimal range of 60% - 80%. For example, the response time of buffer expansion / contraction for the Modbus protocol is shortened from the traditional 2 seconds to 100ms.

[0028] 5. High reliability and scalability at the system level Distributed architecture: The distributed deployment of edge computing nodes reduces the impact range of single - point failures in the system to less than 10%. For example, the failure of an edge node on a production line only affects 3 local devices and does not affect other production lines.

[0029] Docker containerized deployment: The containerized encapsulation of the parsing module enables the system to support the rapid addition of new protocols, and the maintenance cost is reduced by 70%. For example, when adding 5 new device protocols in a factory, there is no need to restart the entire system.

[0030] Closed - loop feedback mechanism: Through the dynamic feedback detection module, realize the full - process optimization of protocol recognition, parsing, and conversion. For example, the iteration cycle of the parsing model for an unknown - protocol device is shortened from the traditional 3 months to 1 week.

[0031] Through the five core innovations of intelligent protocol recognition, standardized parsing, dynamic conversion, adaptive buffering, and closed - loop feedback, this system solves the four major pain points of protocol heterogeneity, real - time requirements, network congestion, and equipment reuse in the industrial Internet of Things, achieving 99.9% communication reliability, 1ms - level latency for real - time protocols, 300% system throughput increase, and 70% maintenance cost reduction, providing an efficient and reliable solution for the heterogeneous network integration of Industry 4.0. Brief description of the drawings

[0032] Figure 1 It is the schematic diagram of the multi - protocol adaptive control system of the industrial computer based on edge computing for the embodiments of this application.

[0033] Figure 2This is the schematic diagram of the reverse characterization tree of a Modbus RTU protocol frame in an embodiment of the present application. Detailed implementation manners

[0034] Next, in combination with the accompanying drawings in the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described. Obviously, the described embodiments are part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present application.

[0035] As Figure 1 shown, the industrial computer multi-protocol adaptive control system based on edge computing includes an industrial computer control center and edge computing nodes. The industrial computer control center is distributedly connected to a number of industrial devices through the edge computing nodes. The edge computing nodes include a protocol recognition module, a protocol parsing module, a protocol conversion module, an adaptive buffer management module, and a dynamic feedback detection module; The protocol recognition module is used to perform CRC matching and address feature Bayesian classification on new industrial devices accessing the industrial network diagram, and set a temporary dynamic feedback detection mechanism or perform lightweight feature matching on the new industrial devices according to the matching results, so as to obtain the communication protocols of the new industrial devices or mark the new industrial devices as in an unknown protocol state; The protocol parsing module is used to construct the reverse characterization tree of each communication protocol, obtain the tree structure features, node attribute features, sequence features, and Docker container images of each communication protocol, perform protocol parsing on the new industrial devices marked as in an unknown protocol state based on the reverse characterization tree, and perform protocol parsing on the new industrial devices with known communication protocols based on the Docker container images; The protocol conversion module is used to collect the real-time status data of the new industrial devices, and perform dynamic protocol conversion on the parsing results of the new industrial devices according to the real-time status data and the preset protocol conversion rules; The adaptive buffer management module is used to set the adaptive buffer area and dynamic priority of the target device according to the communication protocol of the new industrial device; The dynamic feedback detection module is used to perform temporary dynamic feedback detection on the new industrial devices with a temporary dynamic feedback detection mechanism set.

[0036] It should be further noted that in the specific implementation process, the communication connection relationships among several industrial devices are obtained. Regarding the several industrial devices as nodes and the communication connection relationships among the several industrial devices as the connection relationships between the nodes, an industrial network diagram is constructed. Real-time detection of new industrial device access to the industrial network diagram is carried out. The network scanning tool Nmap is used to scan the IP address segments of each industrial device in the industrial network diagram to identify the IP addresses of the existing industrial devices and the IP addresses of the newly accessed industrial devices. When a new industrial device is detected to be accessed to the industrial network diagram, the data generated by the new industrial device is packed into data packets and uploaded to the protocol recognition module.

[0037] It should be further noted that in the specific implementation process, the protocol recognition module performs CRC matching and address feature Bayesian classification on the new industrial devices accessing the industrial network diagram. The process of setting up a temporary dynamic feedback detection mechanism or performing lightweight feature matching on the new industrial devices according to the matching results includes: Construct a public protocol fingerprint library, where the public protocol fingerprint library includes CRC features and address features corresponding to several public protocols. The CRC features include CRC generation polynomials, initial values, check bit lengths, etc., and the address features include address values, address ranges, encoding rules, etc.; Pack the data generated by the new industrial device into data packets, extract features from the data packets to obtain target CRC features and target address features. Input the target CRC features into the public protocol fingerprint library for CRC matching, and screen out the public protocols from the public protocol fingerprint library whose corresponding CRC features are consistent with the target CRC features. Perform address feature Bayesian classification on the public protocols according to the target address features to obtain the posterior probabilities of the target address features belonging to each public protocol; Among them, the specific process of performing address feature Bayesian classification on the public protocols according to the target address features to obtain the probabilities of the target address features belonging to each public protocol includes: Let C: public protocol category, X: target address feature (such as address value, address range, encoding rule, etc.). Convert the target address feature X into a discrete feature vector. For example, address range (4x / 3x / 0x / 1x), register type (holding register / input register), slave address (0x0100 - 0x01FF), PDO mapping address (0x1C10 - 0x1C13). Vectorize the target address feature into , where is a binary feature (such as "whether it belongs to Modbus4x register"), and n represents the number of discrete feature vectors; Since the address features of each public protocol are independent of each other, construct a simplified posterior probability formula: ; Among them, Denote the posterior probability that the target address feature X belongs to the public protocol C, which is the prior probability, representing the occurrence probability of the public protocol C in the industrial network diagram, and is the likelihood, representing the probability that the public protocol C generates the target address feature ; A preset posterior probability determination threshold is used to screen out the public protocol with the highest posterior probability. Compare the posterior probability of the public protocol with the posterior probability determination threshold. If the posterior probability of the public protocol is greater than or equal to the posterior probability determination threshold, label the public protocol as the communication protocol of the new industrial device and set up a temporary dynamic feedback detection mechanism for the new industrial device. If the posterior probability of the public protocol is less than the posterior probability determination threshold, perform lightweight feature matching; CRC matching does not require complex calculations. More than 80% of the protocols in industrial scenarios have unique CRCs, and direct matching can quickly filter out a large number of irrelevant protocols; The computational complexity of Bayesian classification is much lower than that of deep learning but higher than that of direct CRC matching, and it is suitable as the second step of filtering (most protocols have been filtered out by CRC in the first step, and the number of remaining protocols to be processed is small). For example: Suppose there are 100 protocols in the protocol library. After rapid CRC filtering, 10 protocols remain. Bayesian classification only needs to calculate the probabilities of the address features of these 10 protocols, and the total time consumption is <5ms, meeting the industrial real-time requirements. Through the progressive design logic of rapid filtering first and then fine analysis, the balance between efficiency and accuracy is achieved, meeting the dual requirements of industrial scenarios for real-time and robustness.

[0038] If there is no public protocol in the public protocol fingerprint library whose corresponding CRC feature is consistent with the target CRC feature, perform lightweight feature matching.

[0039] It should be further noted that in the specific implementation process, the process of lightweight feature matching includes: Collect several consecutive data packets generated by the new industrial device, perform multi-dimensional feature extraction on the several consecutive data packets to generate a time feature vector, a structure feature vector, and a data content feature vector, and perform feature extraction according to the CRC matching result and the Bayesian classification result of the address features of the new industrial device to obtain protocol fingerprint features; Construct a private protocol recognition model based on lightweight CNN + LSTM, input the time feature vector, the structure feature vector, the data content feature vector, and the protocol fingerprint features into the private protocol recognition model, and obtain the confidence levels of several private protocols according to the output of the private protocol recognition model; Preset a confidence threshold, screen out the private protocol with the highest confidence, compare the confidence of the private protocol with the confidence threshold. If the confidence is greater than or equal to the confidence threshold, mark the private protocol as the communication protocol of the new industrial device. If the confidence is less than the confidence threshold, mark the new industrial device as in a state of unknown protocol.

[0040] Among them, the time feature vector includes the mean frame interval, the standard deviation of the frame interval, the maximum frame interval, and the minimum frame interval; the structural feature vector includes the frame length and the number of fields; the data content feature vector includes the entropy value of the data field, the occurrence frequency of special values (such as 0xFF (Modbus broadcast address), 0x0000 (initial value)), and the protocol fingerprint feature includes the matching degree between the target CRC feature and the CRC features of each public protocol in the public protocol fingerprint library, and the posterior probability of each public protocol; It should be further noted that during the process of obtaining the CRC matching result of the new industrial device and performing feature extraction, a hierarchical scoring method is used to obtain the matching degree between the target CRC feature and the CRC features of each public protocol in the public protocol fingerprint library. For example: assign weights according to the importance of parameters (example weights: polynomial 40%, number of bits 20%, byte order 20%, initial value 15%, check position 5%) to obtain the target CRC feature. Step 1: Match the number of bits. The number of bits must be the same, otherwise it is directly excluded (for example, if the target is a 16-bit CRC and the CAN bus (15-bit) has a matching degree of 0%). Step 2: Match the polynomial. If the polynomials are exactly the same, full marks are obtained. Otherwise, points are deducted according to the Hamming distance (the number of different bits in binary). (For example, if the target polynomial is 0x8005 (binary 1000 0000 0000 0101) and is exactly the same as Modbus, 100% is obtained; it has 6 different bits from 0x1021 (0001 0000 0010 0001) of Zigbee, and the score is 1 - 6 / 16 = 62.5%). Step 3: Match the byte order, initial value, and check position. Byte order: If the high and low order are the same, 100% is obtained, otherwise 0%. Initial value: If the hexadecimal values are the same, 100% is obtained, otherwise points are deducted according to the difference ratio. Check position: If the positions are the same (such as both at the end of the frame), 100% is obtained, otherwise 0%. Step 4: Calculate the total matching degree by weighted calculation. 。

[0041] It should be further noted that in the specific implementation process, the process of constructing the private protocol recognition model includes: Obtain standard data packets of public protocols from data sources such as PLCopen and OPC Foundation. At the same time, cooperate with device manufacturers to collect data packets of variant protocols of several public protocols, label the types of private protocols for the variant protocols of several public protocols, extract multi-dimensional features from the standard data packets of public protocols and the data packets of private protocols, generate time feature vectors, structural feature vectors, and data content feature vectors. Match the data packets of private protocols with the public protocol fingerprint library by CRC and perform Bayesian classification of address features. Obtain protocol fingerprint features according to the CRC matching results and the Bayesian classification results of address features. Use the time feature vectors, structural feature vectors, data content feature vectors, and protocol fingerprint features as the training set and the test set. Input the training set into the private protocol recognition model for training until the loss function is trained stably, and save the model parameters. Test the private protocol recognition model with the test set until it meets the preset requirements, and output the private protocol recognition model; Since protocol recognition needs to capture both static structural features (such as frame header patterns) and dynamic timing features (such as frame interval rules) simultaneously, in this embodiment, a lightweight CNN+LSTM is selected to construct a private protocol recognition model. The specific architecture includes a CNN layer: 3 convolutional blocks to extract local features (such as the pattern of the first 5 bytes); an LSTM layer: 2 layers of bidirectional LSTM to capture timing dependencies (such as the interval rule of 10 consecutive frames); a fully connected layer: prevent overfitting through Dropout and output the probability distribution of more than 200 protocols. Loss function: cross-entropy loss + Focal Loss (to solve class imbalance). During the training process, continuously update the weights through the backpropagation algorithm to make the loss function gradually decrease until it reaches a stable state. When the model training is completed and the parameters are adjusted, use the test set for the final evaluation to obtain the evaluation results of the model. The evaluation results include classification metrics such as accuracy, recall rate, and F1 score. According to the evaluation results on the test set, determine whether the model meets the expected standards. If the requirements are met, save the model parameters and prepare for deployment; if not, it is necessary to return to a previous stage to re-examine issues such as data quality, model structure, or training strategy.

[0042] It should be further noted that in the specific implementation process, the process of the protocol parsing module constructing the reverse representation tree of each communication protocol and obtaining the tree structure features, node attribute features, sequence features, and Docker container images of each communication protocol includes: Build a protocol parsing database, obtain the parsing logics of several communication protocols (including public protocols and private protocols), where the parsing logics include the fixed fields (such as slave address, function code) and variable fields (such as register address, data value) for parsing data packets, etc. Package the parsing logic of each communication protocol into a Docker container image, and the Docker container image contains independent code, dependency libraries, and configuration files; Obtain sample data packets corresponding to several communication protocols; Perform field splitting and hierarchical analysis on the sample data packets of the communication protocols, obtain the nature of each field and the hierarchical relationship between each field in the sample data packets, define different types of nodes according to the nature and hierarchical relationship of the fields, create corresponding instance nodes for each field and set the attributes of the instance nodes (basic field → basic node, set attributes such as type, length, etc. Composite field → composite node, set the list of child nodes. Check field → check node, specify the check algorithm and associated fields). According to the hierarchical relationship between each field, determine the hierarchical relationship between the corresponding instance nodes created for each field. Starting from the root node, connect each instance node according to the hierarchical relationship to generate the reverse representation tree of the communication protocol, and perform feature extraction on the reverse representation tree of the communication protocol to generate tree structure features, node attribute features, and sequence features; Associate the tree structure features, node attribute features, and sequence features of each communication protocol with the Docker container images of each communication protocol and store them in the protocol parsing database.

[0043] Among them, the nature of the field includes type, length, position, semantics, etc.; Field types include numeric fields (judged by unsigned integers, signed integers, floating-point numbers, etc.), enumerated fields (judged by counting the uniqueness of field values), and check fields (verified by algorithms such as CRC16, CRC32), etc.; Field length and position include fixed-length fields (judged by the unchanged position and length in all data packets, such as the slave address of the Modbus protocol (the 1st byte, fixed 1 byte)), variable-length fields (judged by the length field or delimiter); Field semantics is judged by the protocol interaction logic: for example, the correspondence between the request packet and the response packet to determine the use of the field (such as request ID, response data). The hierarchical relationship of fields includes nested structures (parent - child relationships) and parallel relationships. The nested structures include nested relationships based on field positions (for example, a certain data packet structure is [header][data area][checksum], where the data area contains sub - fields [address][value]. Then the data area is the parent node, and address and value are the child nodes), and hierarchical relationships based on field types. Among them, the hierarchical relationships based on field types include composite fields: containing multiple sub - fields. For example, a Modbus request frame consists of [slave address][function code][data][CRC]. The request frame is a composite node, and the others are sub - nodes; parallel fields: having no nested relationship and being sibling sub - nodes belonging to the same parent node. For example, the address and value in the data area; Node types include basic nodes (leaf nodes), composite nodes (non - leaf nodes), root nodes, and checksum nodes Figure 2 It is a schematic diagram of the reverse representation tree of a Modbus RTU protocol frame. The data packet (hexadecimal) of the Modbus RTU protocol frame is: 01 03 00 01 00 02 C4 0B. Slave address (01, 1 byte, unsigned integer → basic node), function code (03, 1 byte, enumeration → basic node), starting register address (00 01, 2 bytes, unsigned integer → basic node), number of registers (00 02, 2 bytes, unsigned integer → basic node), CRC checksum (C4 0B, 2 bytes, CRC16 → checksum node).

[0044] It should be further noted that in the specific implementation process, for new industrial devices marked as in an unknown protocol state, protocol parsing is performed based on the reverse representation tree, and the process of protocol parsing for new industrial devices with known communication protocols based on Docker container images includes: If a new industrial device is marked as in an unknown protocol state, the data packet of the new industrial device is segmented by fields and hierarchically analyzed to construct a reverse representation tree corresponding to the data packet of the new industrial device. Feature extraction is performed on the reverse representation tree to generate tree - structure features, node - attribute features, and sequence features. The tree - structure features, node - attribute features, and sequence features are input into the protocol - parsing database for matching to obtain the representation - tree similarity of each communication protocol in the protocol - parsing database. The communication protocol with the highest representation - tree similarity is selected, and the communication protocol is marked as the communication protocol of the new industrial device. The Docker container image of the communication protocol is downloaded from the protocol - parsing database, and the data packet of the new industrial device is parsed according to the Docker container image to obtain the key data in the data packet. For example, for the Modbus protocol, parsing the data packet according to the Docker container image can accurately extract key data such as register addresses and data values from the data packet, providing a basis for subsequent data processing and control; Among them, the tree structure features include the depth of the tree, the number of nodes, the node hierarchical relationship, etc. The data packet structures of different protocols have different features. For example, the data packets of some protocols may have a deeper nesting level, while those of other protocols are relatively flat; Node attribute features: include the attribute information of instance nodes, such as field type, value range, dependency relationship between fields, etc. For example, a certain protocol may stipulate that the value of a specific field must be within a certain range, or there is a specific calculation relationship between some fields; Sequence features: include the appearance order and arrangement of fields in the data packet. The arrangement order of fields in different protocols may be different, which is also an important basis for judging the protocol type; Input the tree structure features, node attribute features, and sequence features into the protocol parsing database for matching. The calculation formula for obtaining the similarity of the representation trees of each communication protocol in the protocol parsing database is: Preset the tree structure feature T, , where d represents the depth of the tree, n represents the total number of instance nodes, b represents the average branching factor (number of leaf nodes / number of non-leaf nodes), represents the number of instance nodes in the i-th layer, represents the total number of layers; Preset the attribute feature of each node as A, , where t represents the field type (such as Int, String, encoded as a vector by one-hot), l represents the field length (such as the number of bytes), r represents the value range (such as ([0, 255], normalized to 0 - 1), represents whether it is a reserved field (0 or 1), which can be merged into the encoding of t (for example, adding a Reserved type); Preset the field type sequence as S, , where m represents the number of field types; ; ; ; ; Among them, represents the structural similarity of two reverse representation trees and , represents the feature dimension, such as = 5 corresponds to , represents the average attribute similarity corresponding to all instance nodes traversing the two reverse representation trees, represents the number of instance nodes where the two reverse representation trees match, represents the type weight, represents the cosine similarity of vectors used to measure the direction consistency of two attribute vectors (mainly t and r) represents the normalized absolute difference, used to measure the length difference, and the result range is [0, 1] represents the sequence feature similarity represents the field type sequence and the length of the longest common subsequence represents the representation tree similarity between two reverse representation trees 、 and are weight factors If the new industrial device is not marked as the protocol unknown state, obtain the communication protocol of the new industrial device, download the Docker container image of the communication protocol of the new industrial device from the protocol parsing database, and parse the data packets of the new industrial device according to the Docker container image to obtain the key data in the data packets

[0045] When the new industrial device cannot be detected in the industrial network diagram, delete the Docker container image of the communication protocol corresponding to the new industrial device

[0046] It should be further noted that in the specific implementation process, the process of the protocol conversion module collecting the real-time status data of the new industrial device and performing dynamic protocol conversion on the parsing result of the new industrial device according to the real-time status data and the preset protocol conversion rules includes

[0047] Construct an intermediate conversion model, obtain the industrial device communicating with the new industrial device, mark the industrial device as the target device, pre-construct a device information database, and the device information database is used to record the communication protocol types supported by each industrial device, and obtain the communication protocol types supported by the target device according to the device information database Preset several protocol conversion rules, each protocol conversion rule includes a conversion condition and a target protocol, where the conversion conditions include network load, device type, and device status, collect the real-time status data of the new industrial device, and match the real-time status data of the new industrial device with the conversion conditions of each protocol conversion rule If there is a protocol conversion rule whose corresponding conversion condition matches the real-time status data. For example, Rule 1: If the network load < 30% and the device is a PLC, then execute Modbus → MQTT (high bandwidth efficiency); Rule 2: If the device power < 20%, then execute Modbus → CoAP (low power consumption). Obtain the target protocol of the new industrial device according to the protocol conversion rule, and determine whether the communication protocol type supported by the target device includes the target protocol of the new industrial device. If not, perform a two-way protocol conversion operation. If so, input the target protocol and the key data in the data packet into the intermediate conversion model, and according to the intermediate conversion model, output the target protocol representation form of the key data, and transmit the target protocol representation form of the key data to the target device; If there is no protocol conversion rule whose corresponding conversion condition matches the real-time status data, then determine whether the communication protocol types of the new industrial device and the target device are the same. If they are the same, transmit the key data to the target device. If they are different, input the key data in the data packet and the communication protocol of the target device into the intermediate conversion model, and according to the intermediate conversion model, output the communication protocol representation form of the key data and transmit it to the target device.

[0048] Dynamically match the real-time status data with the protocol rules in real time to achieve dynamic balance among transmission efficiency (bandwidth), real-time performance (delay), and device energy consumption; The traditional solution adopts one-to-one mapping: special converters need to be developed for each pair of protocols (such as Modbus → OPC UA, Profinet → EtherCAT). When adding the Nth protocol, N×(the number of existing protocols) types of conversion logics need to be developed. For example, if there are 10 existing protocols and 1 new protocol is added, 10 conversion modules are required, and the complexity increases exponentially, with poor scalability. The intermediate conversion model can unify these differences and make the conversion more general. The intermediate conversion model includes a parsing module and a conversion module. The parser is only responsible for converting the original data into an intermediate format, and the conversion module only needs to process the intermediate format to the target protocol. In this way, each part can be developed and modified independently. For example, when adding a new protocol, only the parsing module and the corresponding conversion rules need to be added, without modifying other parts. Among them, the process of constructing the intermediate conversion model includes: clarifying the communication protocol types to be supported, collecting sample data of different protocols from various data sources, analyzing the characteristics of the sample data, such as data type (numeric, string, boolean, etc.), value range, data frequency, etc., finding out the common data elements existing in different protocols, such as device identifier, data timestamp, physical quantity value, etc., and using these elements as the basis of the intermediate data model; Define attributes for each data element, including data type, length, unit, etc. Select JSON as the data representation format. Based on the relationships between data elements, design a reasonable data hierarchy. For example, place device information, data, and metadata in different levels for each supported protocol; Define the mapping rules from data of each communication protocol to the intermediate data model. These rules describe how to map specific fields in the protocol to the corresponding data elements in the intermediate data model. For example, in the Modbus protocol, register address 40001 corresponds to the "temperature" data element in the intermediate data model; Write code to implement the conversion from different protocol data to the intermediate data model according to the designed mapping rules; Verify the intermediate data model using the collected sample data to ensure that the process of converting from different protocol data to the intermediate data model is correct.

[0049] It should be further noted that in the specific implementation process, the process of performing two-way protocol conversion operations includes: Input the target protocol of the new industrial device and the key data in the data packet into the intermediate conversion model. According to the intermediate conversion model, output the target protocol representation of the key data. Upload the target protocol representation of the key data to the industrial computer control center. According to the industrial computer control center, convert the target protocol representation of the key data into the communication protocol representation of the target device, and then transmit the communication protocol representation of the key data to the target device.

[0050] It should be further noted that in the specific implementation process, the process of the adaptive buffer management module setting the adaptive buffer area and dynamic priority of the target device according to the communication protocol of the new industrial device includes: Since the communication cycles of different protocols vary greatly (such as 10 times per second for Modbus and 1000 times per second for EtherCAT), this frequency difference will cause the data volume generated by the high-speed protocol to far exceed that of the low-speed protocol in the same network, which may exceed the network bandwidth and cause data congestion. To avoid the above problems, this embodiment allocates an adaptive buffer for each protocol and sets a dynamic priority for each independent buffer. For example, when the EtherCAT delay exceeds 2 ms, expand the buffer capacity and lower the priority of Modbus to avoid bandwidth preemption.

[0051] Pre-set the initial capacities of independent buffers for different communication protocols. For example: High-speed real-time types (such as EtherCAT, 1000 times per second): Set a relatively small initial buffer capacity (such as 50KB), but it can be quickly expanded; Low-speed non-real-time types (such as Modbus, 10 times per second): Set a relatively large buffer capacity (such as 200KB), and can handle data with a delay. Allocate independent buffers for each protocol to avoid priority conflicts caused by the mixing of data from different protocols. Allocate a small-capacity but quickly expandable buffer for high-speed protocols for priority processing; Use a large-capacity but shrinkable buffer for low-speed protocols to delay the processing of non-urgent data. If the communication protocol of the target device is inconsistent with that of the new industrial device, set an independent buffer for the target device according to the initial capacity of the independent buffer of the communication protocol of the new industrial device, monitor the protocol metrics of the independent buffer of the target device to obtain the protocol metrics, and the protocol metrics include: Delay: The time from when the data enters the buffer to when it exits the buffer; Queue length: The number of data packets to be processed in the current buffer. Preset the protocol metric threshold range, compare the protocol metrics with the protocol metric threshold range, and if the protocol metrics are not within the protocol metric threshold range, perform adaptive capacity adjustment on the independent buffer; Among them, the specific formula for performing adaptive capacity adjustment on the independent buffer is: ; Among them, represents the new capacity, represents the current capacity, represents the delay threshold, represents the delay, represents the queue length, represents the queue length threshold, and are the weight coefficients of the delay and the queue length respectively, and ; If the target device is simultaneously connected to k (k > 1) new industrial devices for communication, and the communication protocol of the target device is inconsistent with the communication protocols of all k new industrial devices, then set k independent buffers for the target device according to the initial capacities of the independent buffers of the communication protocols of each new industrial device, obtain the protocol metrics of each independent buffer and the protocol real-time level of the communication protocol corresponding to each independent buffer, and obtain the dynamic priority of each independent buffer according to the protocol metrics and the protocol real-time level of each independent buffer.

[0052] Among them, the specific formula for obtaining the dynamic priority of each independent buffer is: ; Among them, represents the dynamic priority, represents the protocol real-time level.

[0053] It should be further noted that, in the specific implementation process, the process of the dynamic feedback detection module performing temporary dynamic feedback detection on a new industrial device with a temporary dynamic feedback detection mechanism includes: Setting a feedback detection period for the new industrial device with a temporary dynamic feedback detection mechanism, collecting the data logs of the new industrial device and the target device within the feedback detection period, statistically analyzing the data logs, and obtaining the interaction metrics between the new industrial device and the target device. The interaction metrics include data loss frequency, data latency, connection establishment time, disconnection and reconnection times, function execution results, etc.; Presetting an interaction metric threshold range. If the interaction metrics of the new industrial device are not within the interaction metric threshold range, then the communication protocol of the new industrial device is used to perform lightweight feature matching on the new industrial device; If the interaction metrics of the new industrial device are within the interaction metric threshold range, then the temporary dynamic feedback detection mechanism of the new industrial device is removed.

[0054] The above embodiments are only used to illustrate the technical method of the present invention and not to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. Industrial computer multi-protocol adaptive control system based on edge computing, characterized in that: It includes an industrial computer control center and an edge computing node, wherein the industrial computer control center and a number of industrial devices are connected in a distributed manner through the edge computing node, and the edge computing node includes a protocol identification module, a protocol parsing module, a protocol conversion module, an adaptive buffer management module and a dynamic feedback detection module; The protocol identification module is used to perform CRC matching and address feature Bayesian classification on new industrial devices connected to the industrial network diagram, set up a temporary dynamic feedback detection mechanism for the new industrial devices or perform lightweight feature matching based on the matching results, obtain the communication protocol of the new industrial devices, or mark the new industrial devices as unknown protocol status; The protocol parsing module is used to construct the reverse representation tree of each communication protocol, obtain the tree structure characteristics, node attribute characteristics, sequence characteristics and Docker container image of each communication protocol, perform protocol parsing based on the reverse representation tree for new industrial equipment marked as unknown protocol status, and perform protocol parsing based on the Docker container image for new industrial equipment with known communication protocols; The protocol conversion module is used to dynamically convert the parsing results of the new industrial equipment according to the real-time status data of the new industrial equipment and the preset protocol conversion rules; The adaptive buffer management module is used to set the adaptive buffer area and dynamic priority of the target device according to the communication protocol of the new industrial device; The dynamic feedback detection module is used to perform temporary dynamic feedback detection on new industrial equipment that is equipped with a temporary dynamic feedback detection mechanism.

2. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 1 is characterized in that: Acquire the communication connection relationship between several industrial devices, take several industrial devices as nodes, and use the communication connection relationship between several industrial devices as the connection relationship between nodes to construct an industrial network diagram, and perform real-time detection of new industrial equipment access to the industrial network diagram. When it is detected that a new industrial device is connected to the industrial network diagram, the data generated by the new industrial device is packaged into a data packet and uploaded to the protocol identification module.

3. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 2 is characterized in that: The protocol identification module performs CRC matching and address feature Bayesian classification on new industrial devices connected to the industrial network diagram. The process of setting a temporary dynamic feedback detection mechanism or performing lightweight feature matching on the new industrial devices according to the matching results includes: Constructing a public protocol fingerprint library, wherein the public protocol fingerprint library includes CRC features and address features corresponding to a number of public protocols; Pack the data generated by the new industrial equipment into data packets, extract features from the data packets, obtain target CRC features and target address features, select public protocols whose corresponding CRC features are consistent with the target CRC features from the public protocol fingerprint library, perform address feature Bayesian classification on the public protocols according to the target address features, and obtain the posterior probability of each public protocol; A posterior probability judgment threshold is preset, and the public protocol with the largest posterior probability is screened out. The posterior probability of the public protocol is compared with the posterior probability judgment threshold. If the posterior probability of the public protocol is greater than or equal to the posterior probability judgment threshold, the public protocol is marked as a communication protocol of the new industrial equipment, and a temporary dynamic feedback detection mechanism is set for the new industrial equipment. If the posterior probability of the public protocol is less than the posterior probability judgment threshold, lightweight feature matching is performed; If there is no public protocol in the public protocol fingerprint library whose corresponding CRC feature is consistent with the target CRC feature, lightweight feature matching is performed.

4. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 3 is characterized in that: The process of lightweight feature matching includes: Collect several continuous data packets generated by new industrial equipment, perform multi-dimensional feature extraction on several continuous data packets, generate time feature vectors, structure feature vectors and data content feature vectors, and perform feature extraction based on the CRC matching results and address feature Bayesian classification results of the new industrial equipment to obtain protocol fingerprint features; Construct a private protocol recognition model, input the time feature vector, structure feature vector, data content feature vector and protocol fingerprint feature into the private protocol recognition model, and output the confidence of several private protocols according to the private protocol recognition model; A confidence threshold is preset to screen out the private protocol with the highest confidence, and the confidence of the private protocol is compared with the confidence threshold. If the confidence is greater than or equal to the confidence threshold, the private protocol is marked as the communication protocol of the new industrial equipment. If the confidence is less than the confidence threshold, the new industrial equipment is marked as a protocol unknown state.

5. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 4 is characterized in that: The protocol parsing module constructs the reverse representation tree of each communication protocol. The process of obtaining the tree structure features, node attribute features, sequence features and Docker container images of each communication protocol includes: Build a protocol parsing database, obtain the parsing logic of several communication protocols, and encapsulate the parsing logic of each communication protocol into a Docker container image; Obtain sample data packets corresponding to several communication protocols; Perform field segmentation and hierarchical analysis on sample data packets of the communication protocol, obtain the properties of each field in the sample data packets and the hierarchical relationship between the fields, define different types of nodes according to the properties and hierarchical relationship of the fields, create corresponding instance nodes for each field and set the attributes of the instance nodes, determine the hierarchical relationship between the instance nodes corresponding to each field according to the hierarchical relationship between the fields, start from the root node, connect the instance nodes according to the hierarchical relationship, generate a reverse representation tree of the communication protocol, perform feature extraction on the reverse representation tree of the communication protocol, and generate tree structure features, node attribute features, and sequence features; The tree structure features, node attribute features and sequence features of each communication protocol are associated with the Docker container image of each communication protocol and stored in the protocol parsing database.

6. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 5 is characterized in that: The process of performing protocol parsing based on the reverse characterization tree for new industrial equipment marked as unknown protocol state and based on the Docker container image for new industrial equipment with known communication protocol includes: If the new industrial equipment is marked as a protocol unknown state, perform field segmentation and hierarchical analysis on the data packet of the new industrial equipment, build a reverse representation tree corresponding to the data packet of the new industrial equipment, extract features from the reverse representation tree, generate tree structure features, node attribute features and sequence features, input the tree structure features, node attribute features and sequence features into the protocol parsing database for matching, obtain the representation tree similarity of each communication protocol in the protocol parsing database, select the communication protocol with the highest representation tree similarity, mark the communication protocol as the communication protocol of the new industrial equipment, download the Docker container image of the communication protocol from the protocol parsing database, parse the data packet of the new industrial equipment according to the Docker container image, and obtain key data in the data packet; If the new industrial equipment is not marked as an unknown protocol state, obtain the communication protocol of the new industrial equipment, download the Docker container image of the communication protocol of the new industrial equipment from the protocol parsing database, parse the data packet of the new industrial equipment according to the Docker container image, and obtain the key data in the data packet.

7. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 6 is characterized in that: The protocol conversion module collects the real-time status data of the new industrial equipment. According to the real-time status data and the preset protocol conversion rules, the process of dynamically converting the parsing results of the new industrial equipment includes: Construct an intermediate conversion model, obtain an industrial device that is connected to the new industrial device for communication, mark the industrial device as a target device, and obtain a communication protocol type supported by the target device; Preset several protocol conversion rules, each of which includes conversion conditions and target protocols, collect real-time status data of new industrial equipment, and match the real-time status data of new industrial equipment with the conversion conditions of each protocol conversion rule; If there is a protocol conversion rule whose corresponding conversion condition matches the real-time status data, the target protocol of the new industrial device is obtained according to the protocol conversion rule, and it is determined whether the communication protocol type supported by the target device includes the target protocol of the new industrial device. If not, a two-way protocol conversion operation is performed. If included, the target protocol and key data in the data packet are input into an intermediate conversion model, and a target protocol representation of the key data is output according to the intermediate conversion model, and the target protocol representation of the key data is transmitted to the target device; If there is no corresponding conversion condition and protocol conversion rule that matches the real-time status data, it is determined whether the communication protocol type of the new industrial equipment is consistent with that of the target equipment. If they are consistent, the key data is transmitted to the target equipment. If they are inconsistent, the key data in the data packet and the communication protocol of the target equipment are input into the intermediate conversion model, and the communication protocol representation of the key data is output according to the intermediate conversion model and transmitted to the target equipment.

8. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 7 is characterized in that: The process of performing a bidirectional protocol conversion operation includes: The target protocol of the new industrial equipment and the key data in the data packet are input into the intermediate conversion model, the target protocol representation of the key data is output according to the intermediate conversion model, the target protocol representation of the key data is uploaded to the industrial computer control center, the target protocol representation of the key data is converted into the communication protocol representation of the target equipment according to the industrial computer control center, and then the communication protocol representation of the key data is transmitted to the target equipment.

9. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 8 is characterized in that: The process of the adaptive buffer management module setting the adaptive buffer area and dynamic priority of the target device according to the communication protocol of the new industrial device includes: Pre-set the initial capacity of the independent buffer of different communication protocols. If the communication protocol of the target device is inconsistent with the communication protocol of the new industrial device, set the independent buffer of the target device according to the initial capacity of the independent buffer of the communication protocol of the new industrial device, monitor the protocol index of the independent buffer of the target device, obtain the protocol index, preset the protocol index threshold range, and if the protocol index is not within the protocol index threshold range, adjust the capacity of the independent buffer adaptively. If there are k new industrial devices communicating with the target device at the same time, and the communication protocol of the target device is inconsistent with the communication protocol of the k new industrial devices, then k independent buffers of the target device are set according to the initial capacity of the independent buffer of the communication protocol of each new industrial device, and the protocol indicators of each independent buffer and the protocol real-time level of the communication protocol corresponding to each independent buffer are obtained. The dynamic priority of each independent buffer is obtained according to the protocol indicators and the protocol real-time level of each independent buffer.

10. The industrial computer multi-protocol adaptive control system based on edge computing according to claim 9 is characterized in that: The process of the dynamic feedback detection module performing temporary dynamic feedback detection on new industrial equipment with a temporary dynamic feedback detection mechanism includes: A feedback detection cycle is set for new industrial equipment that sets a temporary dynamic feedback detection mechanism, and interaction indicators between the new industrial equipment and the target equipment are obtained within the feedback detection cycle; Preset an interaction index threshold interval, if the interaction index of the new industrial equipment is not within the interaction index threshold interval, the communication protocol of the new industrial equipment performs lightweight feature matching on the new industrial equipment; If the interaction index of the new industrial equipment is within the interaction index threshold range, the temporary dynamic feedback detection mechanism of the new industrial equipment is eliminated.

Images