Industrial gateway intelligent communication control system and equipment for edge computing

Abstract

The invention discloses an edge computing-oriented industrial gateway intelligent communication control system and equipment, and relates to the technical field of industrial gateway intelligent communication. The industrial gateway intelligent communication control system for edge computing comprises a data classification module, a data scheduling control module and a communication prediction and control module. According to the invention, the communication data received by the industrial gateway is classified to determine the data source type and the corresponding priority, then whether classification optimization is carried out or not is judged based on the parameters in the classification process to improve the classification accuracy, and the data transmission sequence is scheduled according to the priority corresponding to the communication data. Finally, whether scheduling optimization is carried out or not is judged based on parameters in the scheduling process to improve the scheduling efficiency, prediction is carried out based on the to-be-predicted communication data, network resource allocation is carried out to avoid communication congestion, the control effectiveness of the intelligent communication of the industrial gateway is improved, and the problem that the control effectiveness of the intelligent communication of the industrial gateway is low in the prior art is solved.

Description

Technical Field

[0001] This invention relates to the field of intelligent communication technology for industrial gateways, and in particular to an intelligent communication control system and device for industrial gateways oriented towards edge computing. Background Technology

[0002] The rapid development of the Industrial Internet of Things (IIoT) has led to a significant increase in the number of industrial devices and sensors. However, these devices often employ different communication protocols, resulting in difficulties in data interoperability. Industrial gateways, acting as bridges connecting devices with different protocols, can effectively solve the interconnection problem and achieve efficient data transmission and sharing. Edge computing technology is increasingly being applied to industrial gateways to enable localized processing and real-time response to massive amounts of data from industrial terminal devices. By processing data closer to the data source, edge computing gateways significantly reduce latency and improve system response speed. As a key carrier of edge computing, industrial gateways play a crucial role in real-time data processing and resource scheduling, providing strong technical support for the IIoT.

[0003] In existing technologies, industrial gateways are responsible for collecting data from devices such as sensors and controllers, performing localized processing, and then converting the communication protocols used by different devices to achieve interconnection between devices. Through industrial gateways, users can remotely program, debug, and control devices such as PLCs.

[0004] For example, Chinese patent application CN118612271A discloses a communication method, TCP gateway, and system for a TCP gateway, including: establishing a TCP connection path between the TCP gateway and a third-party device; transmitting target data between the TCP gateway and the third-party device through the TCP connection path; wherein the target data includes data obtained from a target storage space, or the target data includes data stored in the target storage space; the target storage space is the space occupied by multiple logical slave devices managed by the TCP gateway, and each of the multiple logical slave devices corresponds to a third-party device.

[0005] For example, the Chinese patent application CN116389513B discloses a cloud-based multi-robot vision collaborative communication control method and system, which includes: sending and receiving communication information to the control unit through an industrial gateway on the device side, and sending the collected control unit information to the edge subsystem communication node; the active control unit detection module in the edge subsystem communication node generates a sub-thread group with a corresponding active ID based on the received active control unit information, and requests cloud services from the cloud subsystem; the cloud subsystem processes the control unit information through a container based on the received cloud service request from the edge subsystem communication node, and the container management module of the cloud subsystem queries the dynamic ID database at a set frequency and compares it with the ID number of the container currently running in the cloud subsystem, and controls the opening and closing of the container with the corresponding ID based on the comparison result to complete the communication control.

[0006] The above-mentioned technology has at least the following technical problems: In existing technologies, industrial environments typically involve multiple data types, including sensor data, log data, and control commands. These data are transmitted together in the same communication channel, lacking an effective priority management mechanism. This leads to delays in the transmission of critical data (such as control commands) due to bandwidth consumption by non-critical data (such as logs), impacting system real-time performance and stability. Furthermore, when data characteristics are ambiguous or overlapping (e.g., some sensor data and control command identifiers are similar), misjudgments may occur, incorrectly classifying high-priority data as low-priority.

[0007] In terms of bandwidth resource allocation, burst traffic is a common and unavoidable phenomenon in industrial scenarios. This is mainly determined by the dynamic nature of the industrial environment, the complexity of equipment behavior, and the characteristics of the network architecture. It can lead to problems such as uneven bandwidth utilization, latency fluctuations, and degraded communication quality. Traditional industrial gateways mostly adopt static bandwidth allocation methods, which cannot cope with the dynamic changes brought about by burst traffic in industrial scenarios. This easily leads to uneven bandwidth resource utilization, which in turn causes latency fluctuations and reduces communication quality.

[0008] Meanwhile, traditional industrial gateways are mostly limited to data forwarding functions, lacking the ability to intelligently analyze and predict network status and traffic patterns, making it difficult to anticipate network bottlenecks. Their ability to proactively adjust peak loads is limited, hindering dynamic optimization of communication resource allocation. Furthermore, ordinary industrial communication links generally lack encrypted transmission, authentication, and anomaly detection mechanisms, making data vulnerable to eavesdropping, tampering, and other attacks during transmission, posing serious security risks and resulting in low effectiveness of intelligent communication control in industrial gateways. Summary of the Invention

[0009] To address the technical problem of low effectiveness in intelligent communication control of industrial gateways in existing technologies, this invention provides an intelligent communication control system and device for industrial gateways oriented towards edge computing. The technical solution is as follows: On one hand, an intelligent communication control system for industrial gateways oriented towards edge computing is provided. This system is implemented by an intelligent communication control device for industrial gateways oriented towards edge computing. The system includes: a data classification module, a data scheduling control module, and a communication prediction and control module. The data classification module receives communication data from sensors, controllers, and log systems through the industrial gateway, classifies the communication data to determine the data source type and corresponding priority. The data scheduling control module determines whether to perform classification optimization to improve classification accuracy based on parameters during the classification process. If classification optimization is performed, the data transmission order is scheduled according to the priority of the optimized communication data; otherwise, the data transmission order is directly scheduled. The communication prediction and control module determines whether to perform scheduling optimization to improve scheduling efficiency based on parameters during the scheduling process. If scheduling optimization is performed, the network state within the prediction time period is predicted based on the classified communication data to be predicted transmitted to the prediction model after optimization; otherwise, prediction is directly performed based on the communication data to be predicted, and network resources are allocated based on the prediction results to efficiently utilize network resources.

[0010] On the other hand, an intelligent communication control device for an industrial gateway oriented towards edge computing is provided, comprising: a data acquisition device for acquiring communication data from sensors, controllers, and a log system; a processor for classifying the communication data to determine the data source type and corresponding priority, determining whether to perform classification optimization to improve classification accuracy based on parameters in the classification process, scheduling the data transmission order, determining whether to perform scheduling optimization based on parameters in the scheduling process, and making predictions based on the classified communication data to be predicted transmitted to the prediction model; and a memory for storing the communication data, the priority corresponding to the classified communication data, the classification result, the prediction result, and the network resource allocation result. The memory stores computer-readable instructions, which, when executed by the processor, implement any one of the above-described intelligent communication control systems for an industrial gateway oriented towards edge computing.

[0011] The beneficial effects of the technical solutions provided by the embodiments of the present invention include at least the following: 1. By receiving communication data from sensors, controllers, and logging systems through an industrial gateway, centralized data acquisition is achieved, providing a data foundation for subsequent processing. This ensures the acquisition of various key information from the industrial site, classifies and prioritizes the communication data, helps distinguish data of different importance, and provides a basis for subsequent data scheduling and processing. This ensures that important data is processed first, improving the response speed and reliability of the industrial gateway's intelligent communication control. Then, based on the parameters in the classification process, it determines whether to perform classification optimization, dynamically adjusting the classification strategy according to the actual situation to improve classification accuracy. This ensures that communication data is correctly classified and prioritized, providing a more reliable data transmission. The scheduling foundation prioritizes data transmission, ensuring a more rational data transmission order and improving efficiency and timeliness. It then uses parameters from the scheduling process to determine whether optimization is necessary. This allows for dynamic adjustment of scheduling strategies based on actual conditions, enhancing fairness and efficiency, ensuring data of different priorities is transmitted appropriately, and preventing unfairness and resource waste. Furthermore, it predicts network conditions based on the communication data to be predicted, enabling targeted allocation of network resources, proactively mitigating communication congestion, improving network resource utilization, and guaranteeing stable operation and reliable data transmission for the industrial gateway's intelligent communication control.

[0012] 2. By comparing classification time and network resource utilization with reference values, it is possible to determine whether classification optimization is necessary based on the actual situation. This improves the accuracy and efficiency of classification, preventing inaccurate classification from affecting subsequent data processing. If classification optimization is performed, the classification results are checked one by one to ensure accuracy. Abnormal results are marked and manually reviewed to ensure the reliability of data classification. At the same time, the proportion of abnormal classifications is statistically analyzed and alarms are triggered, allowing designated personnel to understand the classification situation in a timely manner and take appropriate measures. If classification optimization is not performed, data transmission order is scheduled. A strict priority queue ensures that the first priority data is processed with absolute priority, meeting the real-time requirements of important data. A weighted fair queue shares bandwidth proportionally to ensure fair transmission of data with different priorities. The transmission rate is adjusted according to network resource utilization to avoid network congestion and improve the operating efficiency of intelligent communication control in industrial gateways.

[0013] 3. Transmission anomaly detection is performed by using queue length and transmission time to promptly identify transmission anomalies, providing a basis for subsequent scheduling optimization and ensuring the stability and reliability of data transmission. Next, a time-slice mechanism is introduced for scheduling optimization, limiting the maximum data volume of a single transmission to prevent first-priority data from occupying bandwidth for extended periods, ensuring that other priority data can also be transmitted in a timely manner. A token bucket rate limiting mechanism is used to control the data transmission rate based on traffic fluctuations, further optimizing network resource utilization. Simultaneously, based on queue length and scheduling waiting time, it determines whether to trigger forced scheduling, enabling timely processing of backlogged data in the weighted fair queue and determining whether to suspend first-priority data transmission to prevent excessive bandwidth occupation. Using bandwidth ensures overall network performance, triggering forced scheduling and pausing of first-priority data transmission according to the scheduling ratio, enabling flexible control of data transmission. While ensuring the transmission of important data, it also considers the transmission needs of other data, improving network fairness and efficiency. Finally, bandwidth optimization is performed based on network resource utilization, rationally allocating bandwidth resources to improve network utilization. The polling frequency is adjusted according to the average queue length fluctuation value to adapt to changes in network traffic, determining whether to allocate minimum bandwidth to ensure the minimum transmission requirements of communication data. Minimum bandwidth is set for communication data of each priority to ensure that communication data can receive basic transmission guarantees and avoid data transmission failure or excessive delay due to insufficient bandwidth.

[0014] 4. By accurately determining the authorization status of IP addresses and the number and frequency of data transmission bytes, potential abnormal access behaviors can be quickly identified, providing accurate evidence for subsequent security handling. If abnormal access behavior is detected, a security alarm mechanism is immediately triggered, and security alarm information is reported to pre-defined personnel, enabling them to quickly understand the security status and providing information support for manual intervention and handling. If no abnormal access behavior is detected, no response measures are taken, maintaining normal operation and avoiding unnecessary interference and resource waste, thus ensuring the operational efficiency of the industrial gateway's intelligent communication control. If there is an abnormal behavior awaiting verification, the pre-defined personnel are notified of the existence of such an abnormal behavior to improve... The system ensures high accuracy and flexibility in identifying abnormal behavior, then implements security response measures to quickly disconnect abnormal communication connections, effectively protecting the security and stability of the industrial gateway's intelligent communication control. It also rationally allocates network resources to prevent network congestion caused by abnormal access. Furthermore, increasing the key length enhances data encryption security and improves the ability to resist password cracking attacks, ensuring the security and integrity of communication data. Finally, after confirming that the abnormal behavior has been resolved, it promptly restores normal operation, reducing the long-term impact of security measures on the performance of the industrial gateway's intelligent communication control. For persistent abnormal behavior, it continuously responds to ensure the effective maintenance of communication data security. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 This is a flowchart of the intelligent communication control of the industrial gateway provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of an industrial gateway intelligent communication control system for edge computing provided in an embodiment of the present invention; Figure 3 This is a flowchart of access anomaly detection provided in an embodiment of the present invention. Detailed Implementation

[0017] This application provides an intelligent communication control system and device for industrial gateways oriented towards edge computing, which solves the problem of low effectiveness of intelligent communication control in existing technologies. By classifying the communication data received by the industrial gateway to determine the data source type and corresponding priority, and then judging whether to optimize the classification based on the parameters in the classification process to improve the classification accuracy, the data transmission order is scheduled according to the priority of the communication data. Finally, the scheduling efficiency is improved based on the parameters in the scheduling process. Based on the communication data to be predicted, prediction is performed and network resources are allocated to avoid communication congestion, thereby improving the effectiveness of intelligent communication control of industrial gateways.

[0018] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0019] like Figure 1The flowchart of the intelligent communication control of the industrial gateway shown below has the following specific logic: Communication data is received and classified to obtain real-time control data (first priority data), monitoring and sensing data (second priority data), and background maintenance data (third priority data). If the classification time is greater than the reference classification time, and the network resource utilization rate reflecting the communication network status is not greater than the reference network resource utilization rate, classification optimization is performed. If the classification time is not greater than the reference classification time, and the network resource utilization rate is not greater than the reference network resource utilization rate, data transmission order scheduling is directly performed. If none of the above conditions are met, transmission anomaly judgment is performed after data transmission order scheduling. If the queue length of each queue is greater than the corresponding reference queue length, and the transmission time of each queue is greater than the corresponding reference transmission duration, the communication data after scheduling optimization is recorded as communication data to be predicted; otherwise, no scheduling optimization is performed, and the communication data after scheduling transmission is directly recorded as communication data to be predicted. Prediction is performed based on the communication data to be predicted, and network resource allocation is performed based on the prediction results. If classification optimization is performed, data transmission order scheduling is performed according to the priority corresponding to the optimized communication data; otherwise, data transmission order scheduling is performed directly.

[0020] This invention provides an intelligent communication control system and device for industrial gateways oriented towards edge computing. For example... Figure 2 The diagram shows a structural schematic of an industrial gateway intelligent communication control system for edge computing. The system includes a data classification module, a data scheduling control module, and a communication prediction and control module.

[0021] The data classification module receives communication data from sensors (such as temperature, pressure, and vibration sensors, sending periodic or event-triggered status monitoring data), controllers (such as PLCs, sending control commands to actuators or receiving configuration commands from the host computer), and log systems (such as operating logs and fault logs generated by the equipment itself or the host computer) through the industrial gateway. This enables centralized data collection, providing comprehensive data support for subsequent processing and ensuring that the system can acquire various key information from the industrial site. The module classifies the communication data to determine the data source type and corresponding priority (a measure of the importance of communication data from different data source types). High-priority communication data receives priority during transmission and processing to ensure that key information is transmitted and processed promptly and accurately. This helps distinguish data of different importance, providing a basis for subsequent data scheduling and processing, enabling key data to be prioritized, and improving the response speed and reliability of the industrial gateway's intelligent communication control.

[0022] The data scheduling and control module determines whether to perform classification optimization to improve classification accuracy based on parameters during the classification process. It can dynamically adjust the classification strategy according to the actual situation, improving classification accuracy and ensuring that communication data can be correctly classified and prioritized, providing a more reliable scheduling basis for data transmission. If classification optimization is performed, the data transmission order is scheduled according to the priority of the optimized communication data. The optimized priority can more accurately reflect the importance and urgency of the data, making the data transmission order more reasonable and improving the efficiency and timeliness of data transmission. If classification optimization is not performed, the data transmission order is scheduled directly. If the classification result meets the requirements, unnecessary classification optimization processes are avoided, improving the processing efficiency of intelligent communication control of the industrial gateway.

[0023] The communication prediction and control module determines whether to perform scheduling optimization based on parameters during the scheduling process. It can dynamically adjust the scheduling strategy according to the actual situation during scheduling to ensure that data of different priorities are transmitted reasonably, avoiding unfairness and resource waste in the data transmission process, thereby improving scheduling fairness and efficiency. If scheduling optimization is performed, the network status within the prediction period is obtained based on the classified communication data to be predicted transmitted to the prediction model after optimization. The optimized scheduling strategy makes data transmission more orderly, and the collected data more accurately reflects the network operation, thereby improving the accuracy of network status prediction and providing a more reliable basis for subsequent network resource allocation. If scheduling optimization is not performed, prediction is directly performed based on the communication data to be predicted, simplifying the prediction process, reducing the additional computational overhead caused by scheduling optimization, improving the real-time performance of intelligent communication control of industrial gateways, and allocating network resources based on prediction results. It can predict the network status in advance and allocate network resources in a targeted manner to avoid communication congestion, thereby making efficient use of network resources.

[0024] It should be added that a Long Short-Term Memory (LSTM) network is used as the prediction model. The communication data to be predicted and historical communication data are used as inputs to the prediction model. The output is the network state prediction results for a future prediction period (e.g., the next 5 minutes), including: the predicted total network traffic, the predicted end-to-end transmission delay, the predicted packet loss rate, etc. Decisions are made based on the prediction results. For example, if it is predicted that the network traffic will exceed 90% of the bandwidth in 5 minutes, it is determined that there is a risk of congestion. The network resource allocation strategy includes bandwidth reservation (reserving fixed bandwidth for first-priority data to ensure that it can be unimpeded under any circumstances), traffic shaping (for second-priority data, rate limiting or smoothing is performed in advance to distribute its peak traffic to network idle periods), and dynamic routing (if it is predicted that a certain network path will be congested, some non-critical data streams are switched to backup paths in advance).

[0025] Furthermore, the communication data is categorized to determine the data source type and corresponding priority. The specific process is as follows: Classifying communication data based on its content characteristics avoids applying a uniform processing method to all data. This allows for the use of more suitable and efficient processing algorithms and strategies for different data types, thereby enhancing the targeting of data management and improving overall data processing efficiency. Furthermore, adding priority tags to communication data based on the classification results enables data to be sorted and processed according to priority in network transmission, data processing systems, and other stages. This ensures that high-priority data (such as real-time control data) is processed first, preventing critical operations from failing due to data congestion or processing delays.

[0026] The classification results of communication data (i.e., data source type and corresponding priority) include real-time control data designated as first priority data, monitoring and sensing data designated as second priority data, and background maintenance data designated as third priority data. Communication data is classified into different priority data. The importance of first priority data, second priority data, and third priority data decreases in that order.

[0027] Data content characteristics include instruction identifiers (codes or symbols used to uniquely identify specific instructions in communication data; different instruction identifiers may correspond to different operations such as device start-up, stop, and acceleration), operation parameters (specific parameter settings related to instruction execution, determining details such as the method and extent of instruction execution; for example, in an instruction to control motor speed, the operation parameter is the specific speed value), numerical data (data represented in numerical form, used to quantify various information, such as physical quantities like temperature, pressure, and speed), device IDs (device identifiers, codes used to uniquely identify each device in the communication network), timestamps, log levels (classification identifiers used to indicate the importance or severity of log information), and descriptive text (textual information describing communication data-related events, states, or operations in natural language). Common log levels include DEBUG, INFO, WARNING, and ERROR, which can help quickly locate and filter log data of different importance.

[0028] It should be added that the extracted data content features are input into a pre-trained machine learning classification model, such as a lightweight gradient boosting decision tree model. This model, while ensuring high classification accuracy, has the advantages of fast inference speed and low resource consumption, making it very suitable for running on resource-constrained industrial gateways. The output communication data category labels are then used. For example, if the data contains specific instruction identifiers that correspond to real-time control operations, and combined with information such as operation parameters and device IDs, the data can be identified as real-time control data and thus classified as first-priority data. If the data mainly reflects the monitoring status of the equipment and includes various monitoring values ​​and device IDs, it can be classified as monitoring and sensing data, i.e., second-priority data. Data related to background maintenance, such as data generated by system upgrades and configuration modifications, is classified as background maintenance data, i.e., third-priority data, based on its characteristics.

[0029] Furthermore, based on the parameters obtained during the classification process, it is determined whether to perform classification optimization. The specific process is as follows: If the classification time (the time spent classifying communication data into data source types) is greater than the reference classification time, and the network resource utilization rate reflecting the communication network status is not greater than the reference network resource utilization rate, it indicates that the classification accuracy of the data source type is low. Classification optimization can help improve the accuracy and efficiency of classification and avoid problems in subsequent data transmission and processing due to inaccurate classification. The network resource utilization rate is the result obtained by weighted coupling of each network resource parameter (including network bandwidth, latency, and packet loss rate) based on the weights of the influence of each network resource parameter on the network status set by preset personnel. The reference classification time, reference network resource utilization rate, and reference abnormal classification ratio are set by preset personnel.

[0030] If the classification time is no greater than the reference classification time and the network resource utilization rate is no greater than the reference network resource utilization rate, then the data transmission order is directly scheduled, reducing unnecessary optimization steps and improving the overall processing efficiency of intelligent communication control of industrial gateways.

[0031] If none of the above conditions are met, a transmission anomaly is determined after the data transmission order is scheduled. Based on the determination result, it is determined whether to optimize the scheduling. This can ensure that the basic data transmission process is normal while timely detection and handling of potential transmission problems, thereby improving the stability and reliability of intelligent communication control of industrial gateways.

[0032] It should be added that by comparing the classification time with the reference classification time, as well as the network resource utilization rate with the reference network resource utilization rate, the rationality of the classification process can be comprehensively evaluated from both time and resource utilization dimensions. This can quickly identify potential problems in the classification process, such as low classification algorithm efficiency or difficulty in classification due to complex data features, thus providing a basis for subsequent optimization decisions.

[0033] The specific process for classification optimization is as follows: Each classification result of communication data is checked to determine whether it conforms to the classification rules (predefined criteria and standards for classifying communication data, such as determining the category of data based on its source, format, and content characteristics; if the data contains a specific instruction identifier, which corresponds to a real-time control operation, and combined with information such as operation parameters and device ID, it can be determined that the data belongs to real-time control data). This ensures that each classification result conforms to the preset classification rules and improves the accuracy of classification.

[0034] If a classification result does not conform to the classification rules, and the abnormal classification ratio of the quantitative communication data classification accuracy is not greater than the reference abnormal classification ratio, the classification result of the communication data is marked as an abnormal classification result, temporarily stored in the processing queue, and a pre-set personnel are prompted to conduct manual review. The number of abnormal classification results is counted. If the abnormal classification ratio of the quantitative communication data classification accuracy is greater than the reference abnormal classification ratio, a classification alarm is triggered, and a classification alarm message including the abnormal classification result and the corresponding data content characteristics is sent to the pre-set personnel. The abnormal classification ratio represents the proportion of abnormal classification results in all communication data classification results, that is, the ratio of the number of abnormal classification results to the total number of communication data classification results. Through the manual review mechanism, the professional knowledge and experience of the pre-set personnel can be used to further judge and process the abnormal classification results, avoiding classification errors caused by the limitations of the automatic classification algorithm. When the abnormal classification ratio exceeds the reference abnormal classification ratio, a classification alarm is triggered, which can promptly detect abnormal situations that occur during the classification process and send classification alarm information to the pre-set personnel, so that relevant personnel can understand the classification problem in a timely manner and take corresponding measures to optimize and improve, thereby improving the accuracy and stability of classification.

[0035] If the data meets the classification rules, the data transmission order is directly scheduled, avoiding unnecessary optimization steps, improving the processing efficiency of intelligent communication control in industrial gateways, and ensuring that data can be transmitted in a timely and accurate manner.

[0036] Furthermore, the data transmission order is scheduled, and the specific process is as follows: First-priority data is assigned to a strict-priority queue to ensure absolute priority processing. Data in low-priority queues is only processed when high-priority queues are empty, thus improving the reliability and real-time performance of first-priority data transmission. Second-priority data is assigned to a second-priority queue, and third-priority data is assigned to a third-priority queue. The weighted fair scheduler extracts and sends data from each weighted fair queue (including the second and third priority queues) according to the set priority weights, ensuring that each weighted fair queue shares bandwidth according to the set priority weights. This allows communication data of different priorities to obtain network resources fairly. By allocating bandwidth through weights, communication data of different priorities can share network resources reasonably according to business needs, avoiding the situation where low-priority communication data is starved due to long periods without transmission opportunities, thus improving the overall utilization and fairness of network resources.

[0037] If the network resource utilization rate is greater than the reference network resource utilization rate, it indicates that the network is in a busy state. The transmission rate of the weighted fair queue is reduced according to the queue reduction ratio until the network resource utilization rate is no greater than the reference network resource utilization rate. Then the transmission rate of the weighted fair queue is restored to maintain the stable operation of the network. By reducing the priority weight of the weighted fair queue, or setting a bandwidth limit in the QoS configuration of the router or switch to reduce the transmission rate, network congestion can be effectively prevented, and problems such as packet loss and increased latency can be avoided, thus ensuring the stability and reliability of the network. By dynamically adjusting the transmission rate through the queue reduction ratio mapping set, bandwidth allocation can be precisely controlled according to the actual network conditions, improving the flexibility of network management.

[0038] The queue reduction ratio represents the result obtained by mapping the network resource utilization rate and the current transmission rate of the weighted fair queue into the queue reduction ratio mapping set. The queue reduction ratio mapping set is a collection obtained from a preset database that represents the mapping relationship between the network resource utilization rate, the current transmission rate of the weighted fair queue, and the queue reduction ratio. The queue reduction ratio mapping set is trained using queue reduction ratio training data, which includes the network resource utilization rate and the current transmission rate of the weighted fair queue based on historical data for historical time periods, as well as the queue reduction ratio set by professional technicians based on experience rules.

[0039] If the network resource utilization rate is not greater than the reference network resource utilization rate, the transmission rate of the weighted fair queue will not be adjusted, and the current transmission status will be maintained. This will make full use of network resources and improve data transmission efficiency.

[0040] Furthermore, the specific process for determining transmission anomalies is as follows: If the queue length (i.e., the amount of data backlogged in the queue) of each queue, including the strict priority queue and the weighted fair queue, is greater than the corresponding reference queue length, and the transmission time of each queue is greater than the corresponding reference transmission duration, then scheduling optimization is performed. Otherwise, scheduling optimization is not performed, and the communication data after scheduling transmission is recorded as the communication data to be predicted, and prediction is made directly based on the communication data to be predicted. The reference queue length, reference transmission duration, and reference traffic fluctuation value are set by preset personnel, and the queue length and transmission time are monitored and recorded in real time by network devices (such as routers, switches, etc.). By comparing the queue length and transmission time with the reference values, it is possible to quickly and accurately determine whether there are abnormalities in network transmission, providing a clear basis for whether to perform subsequent scheduling optimization, helping to adjust network transmission strategies in a timely manner, and ensuring the normal operation of the network.

[0041] Scheduling optimization, specifically: Introducing a time-slicing mechanism into the strict priority queue (dividing a time period into several small, fixed-length time slices, with each queue getting a transmission opportunity within its allotted time slice; once the time slice is used up, other queues take their turn to transmit, thus achieving fair scheduling), after the time slice of the strict priority queue in the current round of scheduling is occupied, the weighted fair queue is scheduled. In the weighted fair queue, the scheduler allocates the remaining bandwidth resources according to the weight ratio of each queue to limit the maximum amount of data transmitted in a single transmission, prevent first-priority data from occupying bandwidth for a long time, and ensure that the weighted fair queue also gets a transmission opportunity. This improves the fairness of network bandwidth resource allocation, ensures that data of different priorities can be transmitted in a timely manner, and improves the overall network performance.

[0042] A time slice represents the result of mapping the queue length and transmission time of each queue into a time slice mapping set. The time slice mapping set is a collection obtained from a pre-set database representing the mapping relationship between queue length, transmission time, and time slices. This set is trained using time slice training data, which includes queue lengths and transmission times from historical data for different time periods, as well as time slices set by experienced technicians based on rules. Dynamically determining the time slice size based on the queue length and transmission time of each queue makes the time slice settings more reasonable, better adapting to the actual transmission needs of different queues. This further improves the flexibility and accuracy of scheduling and optimizes the utilization efficiency of network resources.

[0043] If the traffic fluctuation value, which reflects the degree of traffic fluctuation, is greater than the reference traffic fluctuation value, a token bucket rate limiting mechanism is used to determine whether to trigger forced scheduling. This effectively controls the data transmission rate, prevents network congestion, and ensures network stability. Otherwise, the token bucket limiting mechanism is not used, and a direct determination is made on whether to trigger forced scheduling (under specific conditions, breaking the original scheduling rules and forcibly prioritizing certain queues or data to meet specific network needs or solve urgent problems). The traffic fluctuation value represents the standard deviation of the traffic. By comparing the traffic fluctuation value with the reference value, changes in network traffic can be detected in a timely manner. At the same time, determining whether to trigger forced scheduling based on the traffic fluctuation situation allows for timely response to emergencies in the network and ensures the transmission of critical data.

[0044] The token bucket rate limiting mechanism is adopted, which means setting the token generation rate and preset bucket capacity for the token bucket of the strict priority queue. Data transmission is carried out using tokens. If there are no tokens, data transmission is suspended until a token is generated. The token bucket mechanism can smooth the burst transmission of data, avoid the impact of a large amount of data on the network in a short period of time, effectively prevent network congestion, and ensure the continuity and stability of data transmission.

[0045] The token generation rate represents the result of mapping traffic fluctuation values, queue lengths of each queue, and transmission times of each queue into a mapping set. The token generation rate mapping set is a collection obtained from a preset database that represents the mapping relationship between traffic fluctuation values, queue lengths of each queue, transmission times of each queue, and token generation rates. The token generation rate mapping set is trained using token generation rate training data, which includes traffic fluctuation values, queue lengths, and transmission times of each queue based on historical data for historical time periods, as well as token generation rates set by professional technicians based on empirical rules. This allows the token generation rate to be dynamically adjusted according to the actual network conditions, improving the flexibility and effectiveness of the token bucket rate limiting mechanism.

[0046] Further, determine whether forced scheduling is triggered. The specific process is as follows: If the weighted fair queue length is not less than the queue backlog threshold, or the weighted fair queue scheduling wait time (the length of time data waits to be scheduled) is not less than the scheduling wait time threshold, then forced scheduling is triggered. If forced scheduling is not triggered, it is determined whether to suspend the data transmission of the first priority data. Specifically: if the bandwidth occupied by the strict priority queue is not less than the protection threshold, then the data transmission of the first priority data is suspended according to the scheduling ratio, and the upstream node is notified to reduce the sending rate through protocol signaling (such as RSVP, QoS signaling) until the occupied bandwidth drops below the protection threshold; otherwise, prediction is made directly based on the communication data obtained from transmission. The queue backlog threshold, scheduling wait time threshold, and the protection threshold for the bandwidth occupied by the strict priority queue are set by preset personnel, and the scheduling wait time is monitored and recorded in real time by network devices (such as routers, switches, etc.). By setting protection thresholds, the excessive occupation of network resources by the strict priority queue can be limited. When the bandwidth occupied by the strict priority queue exceeds the protection threshold, its data transmission is suspended according to the scheduling ratio, which can provide bandwidth resources for other data, ensure network fairness and overall performance, and avoid other data streams from being unable to transmit normally due to the excessive occupation of resources by a certain priority data.

[0047] Triggering forced scheduling means increasing the bandwidth of the weighted fair queue according to the scheduling ratio to force the scheduling of communication data in the weighted fair queue until the queue length of the weighted fair queue is less than the queue backlog threshold and the scheduling waiting time of the weighted fair queue is less than the scheduling waiting time threshold. By setting the queue backlog threshold and the scheduling waiting time threshold, it is possible to detect situations where there is too much data backlog or the waiting time in the weighted fair queue in a timely manner, which can speed up the scheduling of data in the queue, avoid data loss due to long-term backlog or network performance degradation, and ensure the real-time performance and reliability of the network.

[0048] The scheduling ratio is obtained by mapping the queue length of the weighted fair queue, the scheduling wait time of the weighted fair queue, and the network resource utilization rate into the scheduling ratio mapping set. The scheduling ratio mapping set is a collection obtained from a preset database that represents the mapping relationship between the queue length of the weighted fair queue, the scheduling wait time of the weighted fair queue, the network resource utilization rate, and the scheduling ratio. The scheduling ratio mapping set is trained using scheduling ratio training data, which includes the queue length of the weighted fair queue, the scheduling wait time of the weighted fair queue, and the network resource utilization rate based on historical data for historical time periods, as well as the scheduling ratio set by professional technicians based on experience rules.

[0049] Further steps include determining transmission anomalies, followed by: If the network resource utilization rate is greater than the reference network resource utilization rate, bandwidth optimization is performed; otherwise, prediction is made directly based on the transmitted communication data. Subsequent operations are determined based on the network resource utilization rate, avoiding unnecessary bandwidth optimization when network resources are sufficient and improving the operating efficiency of intelligent communication control in industrial gateways. When network resources are scarce, bandwidth optimization is performed in a timely manner, ensuring network stability and data transmission quality.

[0050] Bandwidth optimization, specifically: Bandwidth allocation is achieved through network devices (such as routers and switches) configured according to the bandwidth allocation corresponding to each priority data. The bandwidth allocation for each priority data is represented by the bandwidth assigned to that priority data. This allocation is derived from an analysis of the priority weight of each priority data, the data arrival rate (the amount of data arriving at the network per unit time, reflecting the rate at which data flows into the network), and the total bandwidth (the maximum data transmission rate that the network can provide). The priority weight of each priority data is used to reflect the importance of different priority data during bandwidth allocation. The priority weight of each priority data, the reference queue length fluctuation value, and the transmission duration protection threshold are set by pre-defined personnel. Allocating bandwidth according to priority ensures that higher-priority data receives priority access to bandwidth resources, improving the real-time performance and reliability of data transmission, and enhancing the network's service quality.

[0051] If the average queue length fluctuation value within the preset time period is greater than the reference queue length fluctuation value, the polling frequency (the frequency at which data from each queue is processed sequentially according to the scheduling order) is adjusted to the corrected polling frequency, and it is determined whether to allocate minimum bandwidth. Otherwise, the polling frequency is not adjusted, and it is directly determined whether to allocate minimum bandwidth. Adjusting the polling frequency can process the data in the queue more promptly, reduce data waiting time, improve data transmission efficiency, and avoid data loss or transmission delay caused by excessive queue length fluctuations. If the polling frequency is not adjusted, unnecessary frequency adjustment operations are reduced.

[0052] The modified polling frequency representation is the result obtained by mapping the average queue length fluctuation value and the queue length of each queue into the polling frequency mapping set. The polling frequency mapping set is a collection obtained from a preset database that represents the mapping relationship between the average queue length fluctuation value, the queue length of each queue, and the polling frequency. The polling frequency mapping set is trained using polling frequency training data, which includes the average queue length fluctuation value and the queue length of each queue based on historical data for a historical time period, as well as the polling frequency set by professional technicians based on empirical rules. The average queue length fluctuation value represents the average value of all queue length fluctuation values. The queue length fluctuation value is used to reflect the degree of fluctuation of the queue length of each priority data within the preset time period and represents the standard deviation of the queue length of each priority data within the preset time period.

[0053] The determination of whether to allocate minimum bandwidth is as follows: If there is communication data with a transmission duration exceeding the transmission duration protection threshold, then a minimum bandwidth is set for each priority data, and the historical minimum bandwidth of each priority data is recorded as its minimum bandwidth; otherwise, no minimum bandwidth is set for each priority data. When there is communication data with an excessively long transmission duration, setting a minimum bandwidth for each priority data ensures that each priority data has bandwidth resources available for transmission, avoiding long-term data backlog due to insufficient bandwidth, and improving the reliability and stability of data transmission.

[0054] Furthermore, the bandwidth allocation corresponding to each priority level of data is obtained using the following method: The current timestamp and the collection timestamps of each priority data are offset to obtain the collection time difference. If the collection time difference is not less than the reference collection time difference (set by preset personnel), the data importance decay factor is set to the minimum data importance decay factor. Otherwise, the collection time difference of each priority data and the reference collection time difference are normalized to obtain the corresponding data importance decay factor. The data importance decay factor is used to measure the degree to which the importance of data changes over time. For example, the importance of real-time data to decision-making decreases over time. The specific constraint expression of the data importance decay factor is as follows: In the formula, k represents the priority number of the communication data, k=1 represents the first priority data, k=2 represents the second priority data, k=3 represents the third priority data, and δ k δ represents the data importance decay factor corresponding to the k-th priority data, δ0 represents the minimum data importance decay factor, t represents the acquisition time difference, and t0 represents the reference acquisition time difference.

[0055] After averaging the bandwidth utilization rate (i.e., the ratio of actual bandwidth used to total bandwidth) of each priority data within a historical time period, bandwidth occupancy is processed to obtain the corresponding bandwidth occupancy efficiency coefficient. This coefficient reflects the average bandwidth usage of each priority data within the historical time period, reducing ineffective bandwidth allocation. The specific constraint expression for the bandwidth occupancy efficiency coefficient is as follows: In the formula, μ k d represents the bandwidth utilization efficiency coefficient corresponding to the k-th priority data. k This represents the average bandwidth utilization corresponding to the k-th priority data.

[0056] The traffic contribution value for each priority level is obtained by weighting the priority weight of each data level with the corresponding data arrival rate, data importance attenuation factor, and bandwidth utilization efficiency coefficient. In the formula, ω k γ represents the priority weight corresponding to the k-th priority data.k This represents the data arrival rate corresponding to the k-th priority data, and also obtains the traffic contribution ratio of each priority data. This comprehensively measures the role of each priority data in the overall traffic, providing an intuitive proportional basis for the subsequent reasonable allocation of bandwidth based on traffic contribution.

[0057] The bandwidth allocation for each priority data point is obtained by weighting its traffic contribution percentage with the total bandwidth. The specific constraint expression for the bandwidth allocation is as follows: In the formula, B k This represents the bandwidth allocation corresponding to the k-th priority data, where B represents the total bandwidth. This achieves a reasonable allocation of bandwidth based on the contribution of each priority data point to the overall traffic. This makes bandwidth allocation more equitable and reasonable, better meeting the transmission needs of data with different priorities and improving the utilization efficiency of network resources.

[0058] Furthermore, combining industrial internet secure communication protocols, such as message queue telemetry transport protocols, is a lightweight protocol based on a publish / subscribe model, specifically designed for low-bandwidth, high-latency, or unstable networks. Its secure communication is achieved through transport layer security protocols, introducing encryption, authentication, and anomaly detection mechanisms during transmission, such as... Figure 3 The flowchart for access anomaly detection is shown below. The specific process for access anomaly detection is as follows: If the communication Internet Protocol Address (IP address, a digital tag assigned to each device connected to a computer network using the Internet Protocol to communicate, used to uniquely identify the device's location in the network so that data can be accurately sent and received) is an unauthorized address, and the number of bytes in the transmitted data packet exceeds the reference number of bytes, and the transmission frequency exceeds the reference transmission frequency, then abnormal access behavior exists, immediately triggering a security alarm. The security alarm information, including the communication IP address and the transmitted data packet, is fed back to the designated personnel, and security monitoring is conducted after security response measures are taken. The reference number of bytes, reference transmission frequency, bandwidth minimum, initial key length, maximum key length, and security monitoring period are set by the designated personnel.

[0059] If the communication IP address is an authorized address, and the number of bytes in the transmitted data packet is no greater than the reference number of bytes, and the transmission frequency is no greater than the reference transmission frequency, then there is no abnormal access behavior, and no response measures will be taken.

[0060] If neither of the above two conditions is met, a message will be displayed indicating that the preset personnel have engaged in abnormal access behavior that needs to be verified.

[0061] By comprehensively judging the communication IP address, the number of bytes in the transmitted data packets, and the transmission frequency, abnormal access behavior can be identified more accurately, avoiding false positives and false negatives, and improving the security of intelligent communication control of industrial gateways. Once abnormal access behavior is detected, a security alarm is immediately triggered and feedback information is sent to the designated personnel. At the same time, security response measures are taken to quickly prevent the further development of abnormal access and reduce potential security risks. No response measures are taken for cases where there is no abnormal access behavior, avoiding unnecessary consumption of system resources. For abnormal access behavior that needs to be verified, the designated personnel are notified, making security handling more flexible and comprehensive.

[0062] Security response measures include: automatically blocking communication connections for abnormal access behavior, bandwidth limiting, and increasing key length. Bandwidth limiting means adjusting the bandwidth to the lower limit to prevent network congestion; increasing key length means increasing the initial key length to the maximum key length. By automatically blocking communication connections for abnormal access behavior, the transmission of abnormal data is effectively prevented, thus preventing damage to the internal data and equipment of the industrial gateway intelligent communication control system and ensuring the stability of the industrial gateway intelligent communication control system. By limiting bandwidth, network resources are rationally allocated to avoid abnormal access consuming too much bandwidth and causing network congestion, ensuring smooth network communication for other normal services and improving network availability. Strengthening data encryption makes data more difficult to steal and crack during transmission, improving data security and confidentiality.

[0063] Security monitoring specifically involves: if no abnormal access behavior is detected during the security monitoring period, the initial security policy is restored; otherwise, response measures continue to be taken until no abnormal behavior is detected during the security monitoring period. The initial security policy includes the initial key length and the bandwidth before restrictions. This allows for timely detection of whether abnormal access behavior has been completely controlled, ensuring that the system is in a secure state before the initial security policy is restored, preventing the recurrence of security risks, and restoring normal operation mode as soon as possible while ensuring security, thereby reducing the impact of security response measures on the normal operation of the industrial gateway's intelligent communication control.

[0064] This application provides an intelligent communication control device for an industrial gateway oriented towards edge computing, comprising: a data acquisition device for acquiring communication data from sensors, controllers, and a log system; sensors for monitoring various physical quantities (such as temperature, pressure, flow rate, etc.) in the industrial production process; controllers for controlling the operation of industrial equipment; and a log system for recording the operating status and event information of the equipment and system. This system can acquire communication data from multiple different types of data sources to ensure comprehensive information about the industrial production process, providing a rich data foundation for subsequent data processing and analysis.

[0065] The processor classifies communication data to determine the data source type and corresponding priority; determines whether to optimize classification based on parameters during the classification process to improve classification accuracy; schedules data transmission order; determines whether to optimize scheduling based on parameters during the scheduling process; and makes predictions based on the classified communication data to be predicted transmitted to the prediction model. It clarifies the data source type and priority, enabling the processor to process different types of data more effectively, prioritizing important data and improving overall data processing efficiency. Through a classification optimization mechanism, it continuously adjusts and improves the classification algorithm based on classification parameters, improving data classification accuracy, reducing misclassification, and providing reliable data support for subsequent data analysis and decision-making. Through a scheduling optimization mechanism, it dynamically adjusts the scheduling strategy based on network conditions, rationally allocates network resources, improves network transmission efficiency, reduces network congestion, and ensures data transmission stability. The processor makes predictions based on the classified communication data to be predicted transmitted to the prediction model. The prediction model is trained using historical data to learn the inherent patterns and trends of the data, and then makes predictions based on the currently input classified data, providing a basis for industrial production decisions.

[0066] The memory is used to store communication data, the priority of the classified communication data, the classification results, the prediction results, and the network resource allocation results, providing a reference for subsequent data analysis, troubleshooting, performance optimization, etc. The memory stores computer-readable instructions, which, when executed by the processor, implement any of the above-mentioned intelligent communication control system for industrial gateways oriented towards edge computing.

[0067] In summary, this application embodiment achieves centralized data collection by receiving communication data from sensors, controllers, and logging systems through an industrial gateway. This provides a data foundation for subsequent processing, ensuring the acquisition of various key information from the industrial site. Classifying and prioritizing the communication data helps distinguish data of varying importance, providing a basis for subsequent data scheduling and processing. This ensures that important data receives priority processing, improving the response speed and reliability of the industrial gateway's intelligent communication control. Furthermore, based on parameters from the classification process, it determines whether to perform classification optimization, dynamically adjusting the classification strategy according to actual conditions to improve classification accuracy. This ensures that communication data is correctly classified and prioritized, providing a foundation for data transmission. A more reliable scheduling foundation is established, prioritizing data transmission for more rational ordering, improving efficiency and timeliness. Finally, based on parameters during the scheduling process, it determines whether to optimize the scheduling, dynamically adjusting strategies according to actual conditions to enhance fairness and efficiency, ensuring data of different priorities is transmitted appropriately, avoiding unfairness and resource waste. Furthermore, based on predicted communication data, it allows for targeted network resource allocation by anticipating network conditions, mitigating communication congestion, improving network resource utilization, and ensuring stable operation and reliable data transmission for industrial gateway intelligent communication control.

[0068] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0069] This invention is described with reference to flowchart illustrations and / or block diagrams of systems, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0070] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0071] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0072] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of the invention.

[0073] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

[0074] It should also be understood that the memory in the embodiments of the present invention can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).

[0075] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of the present invention are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0076] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0077] In this invention, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.

[0078] It should be understood that, in various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0079] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0080] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the devices, apparatuses, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

Claims

1. An edge computing oriented industrial gateway intelligent communication control system, characterized in that, The system comprises a data classification module, a data scheduling control module and a communication prediction and control module. The data classification module is configured to receive communication data from sensors, controllers and log systems through an industrial gateway, and classify the communication data to determine data source types and corresponding priorities. The data scheduling control module is configured to determine whether to perform classification optimization based on parameters in the classification process to improve classification accuracy. If classification optimization is performed, the data transmission order is scheduled according to the priorities of the communication data after classification optimization.

2. The edge computing oriented industrial gateway intelligent communication control system according to claim 1, wherein, Otherwise, the data transmission order is directly scheduled. The communication prediction and control module is configured to determine whether to perform scheduling optimization based on parameters in the scheduling process to improve scheduling efficiency. If scheduling optimization is performed, the network state in the prediction period is obtained by predicting based on the classified communication data that is transmitted to the prediction model after scheduling optimization. Otherwise, the prediction is directly performed based on the communication data to be predicted, and network resources are allocated based on the prediction results to efficiently utilize network resources. The classification of communication data to determine data source types and corresponding priorities comprises the following steps: Classify the communication data according to data content features to obtain communication data classification results.

3. The edge computing oriented industrial gateway intelligent communication control system according to claim 1, wherein, Add priority labels to the communication data according to the communication data classification results. The data content features include instruction identifiers, operation parameters, numerical data, device IDs, timestamps, log levels and descriptive text. The communication data classification results include real-time control data as first priority data, monitoring and sensing data as second priority data, and background maintenance data as third priority data. The data importance of the first priority data, the second priority data and the third priority data decreases in turn. The determination of whether to perform classification optimization based on parameters in the classification process comprises the following steps: If the classification time is greater than the reference classification time, and the network resource utilization reflecting the communication network state is not greater than the reference network resource utilization, classification optimization is performed. If the classification time is not greater than the reference classification time, and the network resource utilization is not greater than the reference network resource utilization, the data transmission order is directly scheduled. If none of the above conditions is met, transmission abnormality determination is performed after the data transmission order is scheduled, and the scheduling optimization is determined based on the determination result.

4. The edge computing oriented industrial gateway intelligent communication control system according to claim 3, wherein, The specific process of classification optimization is as follows: Check the communication data classification results one by one to determine whether they meet the classification rules. If the classification rules are not met, mark the communication data classification results as classification abnormal results, temporarily store them in the to-be-processed queue, and prompt the preset personnel for manual review. The number of classification abnormal results is counted. If the abnormal classification proportion of the classification accuracy of the quantized communication data is greater than the reference abnormal classification proportion, a classification alarm is triggered, and classification alarm information including the classification abnormal results and corresponding data content features is sent to the preset personnel. If the classification rules are met, the data transmission order is directly scheduled. The specific process of scheduling the data transmission order is as follows: The first priority data is allocated to a strict priority queue to ensure priority processing of the first priority data, and the second priority data and the third priority data are allocated to different weighted fair queues to ensure that each weighted fair queue shares bandwidth according to a set priority weight; If the network resource usage is greater than the reference network resource usage, the transmission rate of the weighted fair queue is reduced according to a queue reduction ratio until the network resource usage is not greater than the reference network resource usage, and the transmission rate of the weighted fair queue is restored; If the network resource usage is not greater than the reference network resource usage, the transmission rate of the weighted fair queue is not adjusted. The queue reduction ratio is a result of mapping the network resource usage, the current transmission rate of the weighted fair queue, and the queue reduction ratio mapping set.

5. The edge computing oriented industrial gateway intelligent communication control system according to claim 3, wherein, The specific process of the transmission anomaly determination is as follows: If the queue length of each queue including the strict priority queue and the weighted fair queue is greater than the corresponding reference queue length, and the transmission time of each queue is greater than the corresponding reference transmission time, scheduling optimization is performed, otherwise, scheduling optimization is not performed, the communication data after scheduling transmission is recorded as predicted communication data, and prediction is directly performed based on the predicted communication data; The scheduling optimization is specifically: A time slice mechanism is introduced in the strict priority queue, after the time slice of the strict priority queue in this round of scheduling is occupied, the weighted fair queue is scheduled to limit the maximum data amount of single transmission and avoid long-time occupation of bandwidth by the first priority data; The time slice is a result of mapping the queue length of each queue and the transmission time of each queue in the time slice mapping set; If the traffic fluctuation value reflecting the fluctuation degree of the traffic is greater than the reference traffic fluctuation value, a token bucket rate limiting mechanism is used to determine whether forced scheduling is triggered, otherwise, forced scheduling is directly determined whether to be triggered; The token bucket rate limiting mechanism means that the token generation rate and the preset bucket capacity of the token bucket are set for the strict priority queue, tokens are used for data transmission, and if there is no token, data transmission is suspended until a token is generated. The token generation rate is a result of mapping the traffic fluctuation value, the queue length of each queue, and the transmission time of each queue in the mapping set.

6. The edge computing oriented industrial gateway intelligent communication control system according to claim 5, wherein, The specific process of determining whether to trigger forced scheduling is as follows: If the queue length of the weighted fair queue is not less than the queue backlog threshold, or the scheduling waiting time of the weighted fair queue is not less than the scheduling waiting time threshold, forced scheduling is triggered, otherwise, whether to suspend data transmission of the first priority data is determined, specifically: if the occupied bandwidth of the strict priority queue is not less than the protection threshold, data transmission of the first priority data is suspended according to a scheduling ratio until the occupied bandwidth decreases to below the protection threshold, otherwise, prediction is directly performed based on the communication data obtained by transmission; The scheduling ratio is a result of mapping the queue length of the weighted fair queue, the scheduling waiting time of the weighted fair queue, and the network resource usage in the scheduling ratio mapping set. The triggering forced scheduling means forcibly scheduling the communication data in the weighted fair queue according to the scheduling proportion until the queue length of the weighted fair queue is less than the queue backlog threshold and the scheduling waiting time of the weighted fair queue is less than the scheduling waiting time threshold.

7. The edge computing oriented industrial gateway intelligent communication control system according to claim 5, wherein, The transmission exception determination further includes: If the network resource usage rate is greater than the reference network resource usage rate, bandwidth optimization is performed, otherwise, the communication data obtained through transmission is directly used for prediction; The bandwidth optimization specifically includes: The bandwidth is allocated according to the bandwidth allocation amount corresponding to each priority data, and the bandwidth allocation amount is obtained based on the priority weight, data arrival rate and total bandwidth of each priority data; If the average queue length fluctuation value in the preset time period is greater than the reference queue length fluctuation value, the polling frequency is adjusted to the corrected polling frequency, and it is determined whether to allocate the minimum bandwidth, otherwise, it is directly determined whether to allocate the minimum bandwidth, the corrected polling frequency is obtained by mapping the average queue length fluctuation value and the queue length of each queue into a polling frequency mapping set, the queue length fluctuation value is used to reflect the fluctuation degree of the queue length of each priority data in the preset time period, and the standard deviation of the queue length of each priority data in the preset time period is represented; The determination of whether to allocate the minimum bandwidth specifically includes: if there is communication data with a transmission time greater than the transmission time protection threshold, setting the minimum bandwidth for each priority data, and recording the historical minimum bandwidth of each priority data as the minimum bandwidth of each priority data.

8. The edge computing oriented industrial gateway intelligent communication control system according to claim 7, wherein, The bandwidth allocation amount corresponding to each priority data is specifically obtained as follows: The current timestamp and the collection timestamp of each priority data are subjected to deviation processing to obtain a collection time difference, if the collection time difference is not less than the reference collection time difference, the data importance decay factor is set to the minimum data importance decay factor, otherwise, the collection time difference and the reference collection time difference of each priority data are subjected to normalization processing to obtain the corresponding data importance decay factor; After the bandwidth utilization rate of each priority data in the historical time period is averaged, bandwidth occupation processing is performed to obtain the corresponding bandwidth occupation efficiency coefficient; The priority weight, the corresponding data arrival rate, the data importance decay factor and the bandwidth occupation efficiency coefficient of each priority data are subjected to weighted processing to obtain the traffic contribution value of each priority, and the traffic contribution proportion of each priority data is obtained; The traffic contribution proportion of each priority data and the total bandwidth are subjected to weighted processing to obtain the bandwidth allocation amount corresponding to each priority data. 9.The edge computing oriented industrial gateway intelligent communication control system according to claim 1, wherein, Further includes: If the communication IP address is an unauthorized address, the number of bytes of the transmission data packet is greater than the reference byte number, and the transmission frequency is greater than the reference transmission frequency, a security alarm is triggered immediately, the security alarm information including the communication IP address and the transmission data packet is fed back to the preset personnel, and security monitoring is performed after taking security response measures; If the communication IP address is an authorized address, the number of bytes of the transmission data packet is not greater than the reference byte number, and the transmission frequency is not greater than the reference transmission frequency, no response measure is taken; If neither of the above two conditions is met, the preset personnel is prompted that there is an abnormal access behavior to be verified. The security response measures include automatically blocking a communication connection of the abnormal access behavior, performing bandwidth limitation, and increasing a key length. The security monitoring specifically includes: if there is no abnormal access behavior in a security monitoring time period, resuming an initial security policy, otherwise, continuing to take response measures until there is no abnormal behavior in the security monitoring time period, the initial security policy including an initial key length and a bandwidth before limitation.

10. An edge computing oriented industrial gateway intelligent communication control device, characterized in that, Comprise: a data acquisition device for acquiring communication data from sensors, controllers and log systems; a processor for classifying the communication data to determine data source types and corresponding priorities, judging whether to perform classification optimization to improve classification accuracy based on parameters in the classification process, scheduling data transmission order, judging whether to perform scheduling optimization based on parameters in the scheduling process, and making predictions based on classified to-be-predicted communication data reflected to a prediction model; a memory for storing communication data, priorities corresponding to classified communication data, classification results, prediction results and network resource allocation results, the memory storing computer readable instructions, the computer readable instructions being executed by the processor to implement the method of any one of claims 1 to 9.

Images