Internet of Things industrial data processing system and method for multi-device access

Abstract

The invention provides an Internet of Things industrial data processing system and method oriented to multi-device access, and the method comprises the steps: obtaining real-time network signal strength and delay indexes from a multi-device access industrial Internet of Things environment, determining a communication link health degree level through preset threshold comparison, and obtaining a link health assessment result for subsequent processing; transmitting a data packet collected by a sensor through the channel allocation scheme, obtaining a packet loss rate index in a transmission process, judging whether the packet loss rate is lower than a preset threshold value, and obtaining a transmission reliability verification result for dynamically adjusting a strategy; and adjusting data transmission parameters according to the measure execution instruction, obtaining an adjusted network throughput index, judging whether the throughput meets a production instruction execution requirement or not, and obtaining an optimization confirmation result for cyclic feedback of link health assessment.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to an Internet of Things (IoT) industrial data processing system and method for multi-device access. Background Technology

[0002] In the field of Industrial Internet of Things (IIoT), building a data processing system that supports multi-device access is particularly crucial. This area directly relates to the efficiency and safety of industrial production, as massive numbers of devices require real-time data transmission to support production decisions and equipment monitoring; its importance is self-evident. With the surge in the number of devices in industrial scenarios and the increasing complexity of data transmission demands, the stability of network communication has become a core pillar for ensuring system operation. Any communication interruption or delay could lead to production stagnation or even safety hazards.

[0003] However, current network communication assurance methods often fall short when dealing with complex industrial environments. Many solutions lack the ability to dynamically perceive network conditions, fail to detect fluctuations in communication quality in a timely manner, and struggle to quickly adjust strategies when problems occur. This deficiency is particularly pronounced in scenarios with multiple devices accessing the network concurrently, especially in industrial settings where network environments are frequently affected by factors such as equipment interference and signal attenuation, making it difficult to guarantee the reliability of data transmission.

[0004] Focusing on the technical challenges, ensuring network communication quality in the Industrial Internet of Things (IIoT) faces two core difficulties. First, there's the issue of real-time assessment of communication link health. Due to frequent changes in network status in industrial environments and the lack of continuous monitoring and judgment mechanisms for communication quality, systems often react passively only after problems occur, unable to provide early warnings or take preventative measures. Second, and related to this, is the insufficient reliability of critical data transmission. Especially when network quality is poor, data loss or delays can directly impact the execution of production instructions. For example, on an automated production line, if sensor data fails to reach the control center in time due to network jitter, it may lead to equipment malfunctions or shutdowns, resulting in economic losses.

[0005] Therefore, how to monitor and maintain the health of communication links in real time in complex industrial environments with multiple devices accessing the system, while ensuring the reliable transmission of critical data, has become a key issue in building an efficient industrial data processing system. Summary of the Invention

[0007] This invention provides an IoT industrial data processing system and method for multi-device access, mainly including: Real-time network signal strength and latency indicators are obtained from an industrial IoT environment with multiple devices connected. The health level of the communication link is determined by comparing it with preset thresholds, and the link health assessment results are used for subsequent processing. The network status changes are determined based on the link health assessment results. If the network status changes exceed a preset threshold, a classification algorithm is used to analyze historical data to determine the communication quality fluctuation pattern. The fluctuation pattern identification results are then used for strategy formulation. Based on the fluctuation pattern recognition results, priority labels are obtained for key data. The risk level of data loss and delay is analyzed by decision tree algorithm to determine whether to activate the backup transmission path. The resulting path activation decision is used to maintain data reliability. If the path activation decision indicates a switch, then the available bandwidth information is obtained from the backup path, the bandwidth resources are grouped using a clustering algorithm, the optimal transmission channel is determined, and a channel allocation scheme is obtained to connect the production instruction execution process. The data packets collected by the sensor are transmitted through the channel allocation scheme, the packet loss rate index during the transmission process is obtained, it is determined whether the packet loss rate is lower than a preset threshold, and the transmission reliability verification result is obtained for dynamic adjustment of the strategy. The real-time monitoring mechanism is updated based on the transmission reliability verification results. If a continuous change in network status is detected, interference source data is obtained from the device logs, interference mitigation measures are determined, and measures execution instructions are obtained to optimize the communication environment. The data transmission parameters are adjusted according to the instructions to implement the measures, the adjusted network throughput index is obtained, it is determined whether the throughput meets the requirements of the production instructions, and the optimization confirmation result is used for the loop feedback of link health assessment.

[0008] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a communication link health assessment and optimization method for industrial IoT environments. It proposes an integrated solution to address complex business problems such as network signal strength, latency fluctuations, data loss, and interference sources in multi-device access scenarios. These problems are interconnected and collectively affect communication quality and data reliability. This invention assesses link health by real-time monitoring of network indicators and comparing them with preset thresholds; it analyzes historical data using classification algorithms to identify fluctuation patterns; it employs a decision tree algorithm to determine whether to activate backup paths and optimizes bandwidth allocation through clustering algorithms; simultaneously, it dynamically adjusts transmission strategies and formulates mitigation measures based on interference source data, ultimately improving network throughput and reliability. The core innovation of this invention lies in forming a closed-loop feedback mechanism for link assessment, pattern recognition, path decision-making, and interference optimization, ensuring efficient execution of production instructions and significantly improving communication stability and data transmission efficiency in industrial IoT environments. Attached Figure Description

[0009] Figure 1 This is a flowchart of an IoT industrial data processing system and method for multi-device access according to the present invention.

[0010] Figure 2 This is a schematic diagram of an IoT industrial data processing system and method for multi-device access according to the present invention.

[0011] Figure 3 This is another schematic diagram of an IoT industrial data processing system and method for multi-device access according to the present invention. Detailed Implementation

[0012] The technical solutions of the embodiments of the present invention will be clearly and thoroughly described below with reference to the accompanying drawings. The described embodiments are merely some embodiments of the present invention.

[0013] like Figures 1-3 This embodiment of an IoT industrial data processing system and method for multi-device access may specifically include: Step S101: Obtain real-time network signal strength and latency indicators from the industrial IoT environment with multiple devices connected, determine the health level of the communication link by comparing with preset thresholds, and obtain the link health assessment result for subsequent processing.

[0014] Real-time network signal strength and latency metrics are continuously collected from each communication link in a multi-device industrial IoT environment. The collected signal strength and latency metrics are compared with corresponding preset thresholds. If the signal strength is below the preset threshold, the link is marked as having abnormal signal strength; if the latency metric exceeds the preset threshold, the link is marked as having abnormal latency. The health level of the communication link is determined by combining these anomaly markers, categorized as healthy, slightly abnormal, or severely abnormal. An isolated forest algorithm is used to detect anomalies in the multi-time-series collected signal strength and latency metrics, correcting misjudgments caused by single-threshold comparisons. The communication link health assessment results are updated based on the corrected anomalies and health levels. A random forest algorithm is used to analyze the correlation between the link health assessment results and historical device operating status, predicting short-term trends in link health levels and generating link health prediction results.

[0015] For example, in an Industrial Internet of Things (IIoT) environment, health assessment and prediction of multi-device communication links can begin with real-time data acquisition and gradually progress to a comprehensive analysis encompassing anomaly detection and trend prediction. First, for real-time network signal strength and latency data collection, assuming a factory has 10 communication links, each link's signal strength and latency data are collected every minute. The preset thresholds for signal strength are -75dBm and latency are 50ms. If a link's signal strength is -80dBm, below the threshold, it is marked as an abnormal signal strength; if the latency is 60ms, above the threshold, it is marked as an abnormal latency. Through combinations of anomaly labels, if a link only has an abnormal signal strength, its health is rated as slightly abnormal; if both are abnormal, it is rated as severely abnormal; if both are normal, it is considered healthy. This grading method intuitively reflects the link status, facilitating subsequent maintenance.

[0016] For example, in anomaly detection, the Isolation Forest algorithm can effectively correct misjudgments caused by comparing single thresholds by analyzing multi-time-series data. Suppose a link's signal strength sequence fluctuates significantly over a period of time, but remains within the normal range overall; a single threshold might misjudge it as an anomaly. However, the Isolation Forest, by constructing random trees, isolates anomaly points from normal points, identifying truly anomalous data points. For instance, a sudden drop in signal strength to -90 dBm at a certain moment, significantly deviating from historical trends, would be identified as an anomaly. This method improves detection accuracy, reduces false alarms, and enhances assessment reliability.

[0017] For example, in the health status update and prediction stage, based on corrected outliers and health levels, a random forest algorithm is used for correlation analysis and trend prediction. Suppose historical data shows that when a link's signal strength is consistently below -80dBm and latency is above 60ms, equipment malfunctions are frequently associated with it. By analyzing the correlation between the current link health status and historical equipment status using the random forest model, it can be predicted that the link may worsen from a minor anomaly to a severe anomaly within the next 24 hours. The prediction results can provide early warnings, guiding maintenance personnel to prioritize checking the link and avoid production losses caused by communication interruptions.

[0018] Specifically, this multi-level analysis method forms a complete closed-loop evaluation system, from data acquisition to anomaly detection and health prediction. The technical choices at each stage, such as the Isolation Forest and Random Forest algorithms, are optimized for the high-dimensional, time-series data characteristics of the Industrial Internet of Things (IIoT) environment, ensuring the accuracy of the evaluation and the practicality of the predictions. This method not only improves the efficiency of communication link management but also significantly reduces the risk of equipment downtime due to communication failures, providing a stable guarantee for industrial production.

[0019] Step S102: Based on the link health assessment results, determine the network status change. If the network status change exceeds a preset threshold, use a classification algorithm to analyze historical data, determine the communication quality fluctuation pattern, and obtain the fluctuation pattern identification result for strategy formulation.

[0020] Real-time evaluation results are acquired through a link health monitoring system. This data undergoes preliminary processing to determine the current network status. Based on this status, the magnitude of status changes is calculated. If the magnitude exceeds a preset threshold, subsequent analysis is triggered to determine the initial range of abnormal fluctuations. For this initial range, corresponding historical data is retrieved from a repository, and a support vector machine algorithm is used to classify the data, resulting in communication quality fluctuation patterns. Based on these patterns, feature data for key time periods is extracted, and it is determined whether these features match known anomaly types. If they do, corresponding pattern recognition labels are generated. These labels are then matched against a pre-established policy library to obtain adjustment schemes related to the labels, determining the specific direction of policy adjustments. After obtaining the specific direction of policy adjustments, corresponding control commands are generated and sent to the network management system to complete the communication quality optimization process. The optimized network status data is used to continuously monitor link health performance, update the evaluation result database, and maintain the accuracy of real-time data.

[0021] For example, in a multi-device environment of the Industrial Internet of Things (IIoT), the link health monitoring system can acquire evaluation results data in real time. This data mainly includes signal strength, latency indicators, and health level classification labels.

[0022] Specifically, the preliminary processing stage involves cleaning and normalizing the collected data to ensure that noise interference is removed and the current network performance is obtained, such as the current average signal strength of -65dBm and the average latency of 45ms.

[0023] In one possible implementation, the magnitude of the state change is calculated based on the current performance of the network. For example, compared to the previous cycle, the signal strength has decreased by 8dB and the latency has increased by 12ms. If this magnitude exceeds a preset threshold, such as a signal strength change exceeding 5dB or a latency change exceeding 10ms, the subsequent analysis process is immediately triggered to quickly determine the initial range of abnormal fluctuations, avoiding unnecessary in-depth processing caused by small fluctuations, thereby improving system response efficiency.

[0024] It should be noted that, for the initial assessment of the abnormal fluctuations, historical data for the corresponding time period was retrieved from the repository. This historical data covered similar link records from the past 30 days. A support vector machine algorithm was used to classify the current and historical data, separating normal and abnormal samples using a hyperplane to obtain the fluctuation patterns in communication quality, such as periodic interference patterns or sudden attenuation patterns. This classification helps to accurately identify the root cause of the fluctuations and improve the accuracy of anomaly localization.

[0025] For example, extracting feature data for key time periods may include peak delay, minimum intensity, and fluctuation frequency.

[0026] In one embodiment, if the feature data shows that the delay repeatedly exceeds 80ms and the intensity is below -75dBm during a specific period, it is determined to conform to a known type of interference anomaly, and a corresponding pattern recognition label, such as "electromagnetic interference type," is generated. This step can effectively reduce misjudgments, classify similar fluctuations into the same problem, and facilitate subsequent unified processing.

[0027] Specifically, by matching pattern recognition tags with a pre-established policy library—for example, the "electromagnetic interference" tag corresponds to schemes such as increasing channel switching frequencies or activating backup links—the specific direction of adjustment can be obtained, such as prioritizing switching to the 5GHz band. This matching mechanism ensures that the adjustment scheme is highly targeted and significantly shortens the optimization response time.

[0028] In one possible implementation, after obtaining the policy adjustment direction, corresponding control commands are generated, such as adjusting router power or reallocating bandwidth, and sent to the network management system for execution. After optimization, network status data will show that the signal strength has recovered to -60dBm, the latency has decreased to 30ms, and the communication quality has been significantly improved.

[0029] For example, continuous monitoring of the optimized link health performance and updating the database with new assessment results ensures the accuracy and continuity of real-time data. This closed-loop monitoring can promptly detect the decay of optimization effects, prevent secondary anomalies, and ultimately maintain the long-term stable operation of communication links in the industrial IoT environment, improving the reliability of device interconnection and production efficiency.

[0030] Step S103: Obtain priority labels for key data based on the fluctuation pattern recognition results, analyze the risk level of data loss and delay using a decision tree algorithm, determine whether to activate the backup transmission path, and obtain a path activation decision to maintain data reliability.

[0031] Obtain the fluctuation pattern identification results. Analyze the fluctuation pattern identification results using a decision tree to obtain a risk assessment. If the risk assessment indicates a risk of data loss, determine whether to activate the backup path. Process the risk assessment using priority labels to obtain label classifications. Determine the path activation decision based on the label classifications. Obtain transmission control commands based on the path activation decision. Perform data reliability maintenance using the transmission control commands.

[0032] In the field of network communication quality optimization, after obtaining the results of fluctuation pattern identification, risk assessment is obtained by analyzing the results through decision tree analysis. For example...

[0033] In one embodiment, the decision tree algorithm uses previously classified fluctuation pattern labels, such as intermittent packet loss or sudden increase in delay, as input nodes for branching decisions.

[0034] Specifically, the decision tree first examines the duration and magnitude of the fluctuation pattern. If the duration exceeds 30 minutes and the magnitude reaches 20%, it branches to a high-risk path, thus outputting a risk assessment of potential data loss risk. This analysis process emphasizes the hierarchical decision-making advantage of decision trees, enabling them to quickly distinguish between normal fluctuations and abnormal risks.

[0035] It should be noted that if the risk assessment indicates a risk of loss, a decision should be made immediately as to whether to activate the backup path.

[0036] In one possible implementation, the system compares the current packet loss rate of the primary path (e.g., reaching 5%) with the health indicators of the backup path (e.g., a packet loss rate of only 1%). If the risk of the primary path is more than twice the threshold of the backup path, then the backup path is activated. This step helps to avoid data interruptions in advance and ensure transmission continuity.

[0037] For example, priority labeling is used to process risk assessments to obtain label classifications.

[0038] Understandably, risk assessment results are assigned priority labels, such as "urgent loss risk" for high priority and "delay risk" for medium priority. The assessment is quantified into labels through rule mapping. This label classification facilitates accurate subsequent decision-making and avoids delays caused by fuzzy processing.

[0039] Specifically, path activation decisions are determined based on label classification.

[0040] In one embodiment, if the tag is of high priority and at high risk of loss, the decision is to directly activate the low-latency backup path; if it is of medium priority, only partial traffic switching is performed. This allows for a flexible balance between load and reliability, reducing unnecessary path switching overhead.

[0041] For example, after determining the path activation decision, a transmission control command is obtained.

[0042] Preferably, the decision result is converted into specific instructions, such as "switch to path B and monitor the traffic allocation ratio as 70:30". The instructions include the target path ID and adjustment parameters to ensure accurate execution.

[0043] In one possible implementation, data reliability maintenance is performed via transmission control commands.

[0044] For example, after the command is sent to the router, the system monitors the packet loss rate of the new path in real time, reducing it to 0.5%, and activates a redundancy verification mechanism to further strengthen data integrity. This maintenance process significantly improves overall communication stability and reduces the probability of service interruptions caused by fluctuations. Through these chained processes, a closed-loop optimization is formed from risk assessment to final maintenance, ensuring the network's adaptability in complex environments.

[0045] Step S104: If the path activation decision indicates switching, obtain available bandwidth information from the backup path, use a clustering algorithm to group the bandwidth resources, determine the optimal transmission channel, and obtain a channel allocation scheme for connecting the production instruction execution process.

[0046] If the path activation decision indicates a switch, available bandwidth information is obtained from the backup path. The current load ratio of each backup path is calculated based on the available bandwidth information. If the load ratio is lower than a preset threshold, the corresponding backup path is marked as a candidate path. A bandwidth resource sequence is obtained from the candidate paths. The bandwidth resource sequence is grouped using a clustering algorithm. The group with the highest bandwidth stability within each group is determined based on the clustering results. The transmission channel with the largest bandwidth peak value is selected as the optimal channel from the group with the highest stability. A channel allocation scheme is generated based on the optimal channel to connect the production instruction execution flow.

[0047] For example, in the field of data transmission reliability maintenance, if a path activation decision indicates a switch to a backup path, the first step is to obtain available bandwidth information from the backup path.

[0048] Specifically, the real-time monitoring module can periodically query the current bandwidth usage of each backup path to obtain data including total bandwidth and used bandwidth.

[0049] In one possible implementation, assuming the system detects a risk on the primary path, it immediately collects data from three backup paths, obtaining that path A has an available bandwidth of 800Mbps, path B has 600Mbps, and path C has 950Mbps. Based on the available bandwidth information, the current load ratio of each backup path is calculated.

[0050] Understandably, the load ratio is calculated by dividing the used bandwidth by the total bandwidth and is used to reflect the degree of congestion on the path.

[0051] For example, continuing the scenario above, path A has a total bandwidth of 1000Mbps, with 200Mbps already used, resulting in a load ratio of 0.2; path B has a total bandwidth of 800Mbps, with 200Mbps already used, resulting in a load ratio of 0.25; and path C has a total bandwidth of 1200Mbps, with 250Mbps already used, resulting in a load ratio of approximately 0.21. This calculation can quickly identify relatively idle paths, helping to avoid switching to already heavily loaded backup paths, thereby reducing the risk of further latency. If the load ratio is below a preset threshold, the corresponding backup path is marked as a candidate path.

[0052] In one embodiment, the preset threshold is preferably set to 0.3 to ensure sufficient margin.

[0053] For example, in the above calculation, the load ratios of all three paths are below 0.3, therefore all are marked as candidate paths. This marking mechanism helps to filter out reliable options and improve transmission stability after switching. A bandwidth resource sequence is then obtained from the candidate paths.

[0054] Specifically, bandwidth sampling points over a period of time can be collected to form a sequence. For example, the sequence for path A is the bandwidth values ​​for the most recent 10 minutes: 780, 810, 795, 820 Mbps, etc. This sequence reflects the dynamic changes of the path, facilitating subsequent analysis of fluctuation trends. Clustering algorithms are used to group the bandwidth resource sequences.

[0055] For example, K-means clustering can be used to divide sequences into high-stability and low-stability groups, and stability can be assessed by calculating the within-group variance.

[0056] In one possible implementation, the path A sequence is clustered into two groups: one group with small variance indicating stable fluctuations, and the other group with large variance indicating occasional peaks and troughs. This grouping effectively separates periods of consistent performance, facilitating the accurate selection of high-quality resources. The clustering results are then used to determine the group with the highest bandwidth stability within each group.

[0057] Understandably, the most stable groups usually have the smallest variance and the highest average bandwidth.

[0058] For example, path A has a high stability group variance of only 15 Mbps and an average bandwidth of 805 Mbps, while the low stability group variance reaches 80 Mbps; therefore, the high stability group is selected. This selection method emphasizes the need for long-term reliable transmission and avoids instability caused by short-term fluctuations. The transmission channel with the largest peak bandwidth is selected from the highest stability group as the optimal channel.

[0059] In one embodiment, among multiple sampling channels within the high-stability group, the channel with a peak speed of 850 Mbps is selected as the optimal one. This selection balances stability and peak capacity, providing better support during periods of increased demand and improving overall data reliability. A channel allocation scheme is then generated based on the optimal channel to connect the production instruction execution process.

[0060] For example, the generated solution specifies the selected channel using path A and issues instructions to the production system to execute the connection switch. This solution generation ensures seamless integration, significantly reduces the probability of data loss, and optimizes resource utilization efficiency. Through the above process, the system can quickly switch to a high-quality backup channel when the primary path fails, maintaining highly reliable transmission and effectively supporting the real-time requirements of industrial data.

[0061] Step S105: Transmit the data packets collected by the sensor through the channel allocation scheme, obtain the packet loss rate index during the transmission process, determine whether the packet loss rate is lower than the preset threshold, and obtain the transmission reliability verification result for dynamic adjustment strategy.

[0062] Sensor data packets are transmitted using a channel allocation scheme, and the sending and receiving acknowledgment timestamps for each packet are recorded. Transmission delay is calculated from these timestamps, and the number of packets without acknowledgment is counted. The packet loss rate is calculated based on the number of unacknowledgmented packets and the total number of transmitted packets. The packet loss rate is then checked against a preset threshold; if it is, the transmission reliability is determined to be high. Channel occupancy and interference intensity are collected under the current channel allocation scheme based on the transmission reliability results. A random forest model is trained using the channel occupancy and interference intensity to obtain channel quality prediction scores. All available channels are ranked according to their prediction scores, and the channel allocation scheme is adjusted to prioritize channels with higher scores.

[0063] For example, after the optimal transmission channel is determined by the channel allocation scheme, the sensor data packets can be transmitted through this scheme, enabling real-time monitoring of the data flow stability during the execution of industrial production instructions.

[0064] Specifically, the system records the sending timestamp and receiving acknowledgment timestamp for each sent sensor data packet to calculate the transmission delay.

[0065] In one embodiment, assuming a data packet is sent at 10:00:00.123 and received at 10:00:00.456, the transmission delay is approximately 0.333 seconds. This low latency helps ensure timely response to production instructions.

[0066] It is understandable that the number of data packets that have not been acknowledged is counted at the same time, and the packet loss rate is calculated based on the total number of data packets sent.

[0067] For example, if a total of 1000 data packets are sent, and 5 packets fail to receive acknowledgments, the packet loss rate is 0.5%. The system checks if this rate is below a preset threshold; if the threshold is set to 1%, then a transmission reliability result below this threshold is considered high reliability. This high reliability assessment directly supports subsequent channel optimization, preventing production interruptions caused by continuous transmission on unreliable paths.

[0068] In one possible implementation, the channel occupancy rate and interference intensity under the current channel allocation scheme are collected based on the transmission reliability results.

[0069] For example, the channel occupancy rate might reach 65%, and the interference intensity might be at a moderate level. By training a random forest model using these features, it is possible to synthesize the channel quality prediction scores from multiple decision trees.

[0070] Preferably, the model will assign higher scores to channels with low occupancy and weak interference, such as one channel scoring 8.7 points while another channel scores only 6.2 points.

[0071] For example, after sorting all available channels according to their predicted channel quality scores, the channel allocation scheme is adjusted to prioritize channels with higher scores. This adjustment allows the system to automatically switch to the better path in the next round of transmission, improving overall data transmission efficiency.

[0072] In one embodiment, the original scheme used a channel with a score of 6.2, which occasionally caused increased latency. After adjustment, an 8.7-score channel was selected, which reduced the average transmission latency by 20% and further reduced the packet loss rate to 0.2%. This significantly improved the stability and real-time performance of the production instruction execution process, ensured that sensor data packets reliably arrived at the control end, and supported accurate production decisions.

[0073] It should be noted that the entire process is closely integrated with the initial path switching and clustering selection of the optimal channel, forming a closed-loop optimization. Higher reliability leads to more accurate model training, further enhancing the intelligence of the channel allocation scheme. This prediction and adjustment based on measured metrics effectively reduces interference, maintains low packet loss rates and high reliability, resulting in consistently stable transmission performance and improving resource utilization efficiency in industrial production environments.

[0074] Step S106: Update the real-time monitoring mechanism based on the transmission reliability verification result. If a continuous change in network status is detected, obtain interference source data from the device log, determine interference mitigation measures, and obtain measure execution instructions to optimize the communication environment.

[0075] The current transmission reliability verification result is obtained through a real-time monitoring mechanism. If the transmission reliability verification result is lower than a preset threshold, it is determined whether the network status is continuously changing. When the network status is continuously changing, interference source data is extracted from the device logs. Based on the extracted interference source data, the interference type is determined to be either co-channel interference or multipath interference. If it is determined to be co-channel interference, frequency switching is used as a mitigation measure. If it is determined to be multipath interference, equalizer parameter adjustment is used as a mitigation measure. Based on the determined mitigation measures, corresponding execution instructions are generated and sent to the communication module.

[0076] For example.

[0077] In one embodiment, the current transmission reliability verification result is continuously obtained through a real-time monitoring mechanism, which is derived from the previous verification process based on packet loss rate and transmission delay.

[0078] Specifically, when the transmission reliability verification results show that the packet loss rate is higher than a preset threshold such as 5%, the system immediately determines whether the network status is undergoing continuous changes.

[0079] In one possible implementation, continuous changes in network status can be confirmed by continuously collecting channel occupancy fluctuations over multiple time windows. For example, if the channel occupancy rate rises from 30% to 70% within the last 10 minutes and remains high, it is determined to be a continuous change. At this point, interference source data is extracted from the device logs, which record the surrounding signal strength, frequency band occupancy, and timestamps of abnormal events.

[0080] It should be noted that the type of interference is further determined based on the extracted interference source data.

[0081] Specifically, if the log shows that multiple strong signal sources are continuously occupying the same frequency band and the signal strength exceeds -70dBm, it is judged as co-channel interference; conversely, if the signal strength fluctuates drastically but the frequency band is singular, and is accompanied by multipath delay extension exceeding 50 microseconds, it is judged as multipath interference.

[0082] For example, in the case of co-channel interference, the system determines to adopt frequency switching as a mitigation measure.

[0083] For example, after the original channel suffers from co-channel interference in the 2.4GHz band, it quickly switches to the idle 5GHz sub-channel, thereby reducing the packet loss rate of subsequent data packet transmission from 8% to below 2%, improving the continuous reliability of sensor data.

[0084] In one embodiment, for multipath interference scenarios, it is preferable to adopt mitigation measures by adjusting equalizer parameters.

[0085] Specifically, by increasing the number of equalizer taps or optimizing the step size of the adaptive algorithm, the receiver can better compensate for inter-symbol interference caused by multipath, thereby significantly reducing the bit error rate and stabilizing the transmission delay to within 100 milliseconds.

[0086] Understandably, based on the determined mitigation measures, the system generates corresponding implementation instructions and sends them to the communication module.

[0087] For example, in the event of co-channel interference, a command to "switch to channel 12" is generated; in the event of multipath interference, a command to "adjust the equalizer coefficient to 0.02" is generated. After these commands are issued, the communication module executes them immediately, ensuring that the data packets collected by the sensor maintain high-reliability transmission in the dynamic network environment, ultimately supporting further optimization and adjustment of the channel allocation scheme.

[0088] Step S107: Execute instructions to adjust data transmission parameters according to the measures, obtain the adjusted network throughput index, determine whether the throughput meets the production instruction execution requirements, and obtain optimization confirmation results for loop feedback in link health assessment.

[0089] For the data transmission process, a pre-established monitoring module acquires current network throughput data to determine initial baseline values ​​for transmission parameters. Based on these baseline values, automated tools simulate and calculate parameter adjustment schemes to obtain adjusted parameter configuration schemes. For the adjusted parameter configuration scheme, this scheme is loaded into a network transmission simulation environment, and the throughput data changes after simulation are observed. If the throughput data is detected to be lower than a preset execution standard, a parameter readjustment process is triggered to determine a new parameter configuration scheme. For the new parameter configuration scheme, it is deployed in an actual network environment, and real-time throughput data is acquired to determine if it meets the execution standards of production instructions. Based on the real-time throughput data, if the execution standards are met, the parameter configuration scheme is recorded in the link health assessment database to obtain a preliminary assessment result of the link health status. Based on the preliminary assessment result of the link health status, a feedback mechanism transmits the assessment cycle data to the parameter adjustment module to determine the optimization direction for the next data transmission.

[0090] In the field of data transmission optimization, by acquiring current network throughput data through pre-established monitoring modules, it is possible to capture link performance fluctuations in real time, thereby providing a reliable basis for subsequent parameter adjustments. For example...

[0091] In one possible implementation, the monitoring module collects throughput data every 5 seconds. When the average throughput is only 150Mbps after 10 consecutive collections, the initial transmission parameter baseline value can be determined to be 150Mbps, which helps to quickly identify performance bottlenecks.

[0092] Specifically, by using automated tools to simulate and calculate the parameter adjustment scheme based on the initial transmission parameter baseline values, the risks associated with trial and error in the production environment can be avoided.

[0093] In one embodiment, the automated tool first inputs baseline values ​​into the simulation model, generates multiple adjustment combinations for parameters such as window size and congestion control algorithm, and after rapid iterative calculation, outputs a set of adjusted parameter configurations that increase the window size by 20%, thereby increasing the simulated throughput to 220 Mbps. This simulation calculation method significantly reduces uncertainty before actual deployment.

[0094] For example, by loading the adjusted parameter configuration scheme into a simulated environment via network transmission, its potential effects can be visually verified.

[0095] It should be noted that the simulation environment reproduces the latency, packet loss, and other characteristics of a real network. After loading a new solution, if the simulated throughput is observed to be stable above 210Mbps, it indicates that the solution is worth further practical verification; otherwise, a readjustment process is triggered. This pre-verification mechanism helps to accurately select effective solutions.

[0096] In one possible implementation, if the throughput data after simulation shows a decrease below the preset execution standard, such as consistently falling below 180Mbps, the parameter readjustment process is immediately triggered. At this point, historical logs can be used to further optimize the congestion control threshold, generating a new parameter configuration scheme that includes reducing retransmission timeout, thus providing a more robust option for actual deployment.

[0097] Specifically, for the new parameter configuration scheme, deployment in a real network environment and acquisition of real-time throughput data can ultimately confirm its production availability.

[0098] For example, if post-deployment monitoring shows a stable throughput of 230Mbps with fluctuations of less than 5%, it is determined to meet the execution standards of production instructions. This practical verification process ensures the reliability of the optimization solution.

[0099] For example, if the real-time throughput data indicates that the performance meets the execution criteria, the parameter configuration scheme is recorded in the link health assessment database, forming a preliminary assessment result of the link health status. This step provides a data foundation for long-term performance tracing and helps accumulate experience to support subsequent optimizations.

[0100] In one embodiment, based on the preliminary assessment results of the link health status, the assessment loop data is transmitted to the parameter adjustment module through a feedback mechanism, which can achieve closed-loop optimization.

[0101] Specifically, the feedback data includes the throughput improvement and stability metrics. In the next optimization, parameters related to the weak points can be prioritized for adjustment, such as further refining the dynamic window growth strategy for latency-sensitive scenarios. This feedback loop continuously improves the overall efficiency and adaptability of data transmission, ensuring that the link maintains a high throughput level under complex network conditions.

[0102] The above description is merely an example and illustration of the structure of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the structure of the invention or exceed the scope defined in the claims, all of which should fall within the protection scope of the present invention.

Claims

1. An IoT industrial data processing system and method for multi-device access, characterized in that, The method includes: Real-time network signal strength and latency indicators are obtained from an industrial IoT environment with multiple devices connected. The health level of the communication link is determined by comparing it with preset thresholds, and the link health assessment results are used for subsequent processing. The network status changes are determined based on the link health assessment results. If the network status changes exceed a preset threshold, a classification algorithm is used to analyze historical data to determine the communication quality fluctuation pattern. The fluctuation pattern identification results are then used for strategy formulation. Based on the fluctuation pattern recognition results, priority labels are obtained for key data. The risk level of data loss and delay is analyzed by decision tree algorithm to determine whether to activate the backup transmission path. The resulting path activation decision is used to maintain data reliability. If the path activation decision indicates a switch, then the available bandwidth information is obtained from the backup path, the bandwidth resources are grouped using a clustering algorithm, the optimal transmission channel is determined, and a channel allocation scheme is obtained to connect the production instruction execution process. The data packets collected by the sensor are transmitted through the channel allocation scheme, the packet loss rate index during the transmission process is obtained, it is determined whether the packet loss rate is lower than a preset threshold, and the transmission reliability verification result is obtained for dynamic adjustment of the strategy. The real-time monitoring mechanism is updated based on the transmission reliability verification results. If a continuous change in network status is detected, interference source data is obtained from the device logs, interference mitigation measures are determined, and measures execution instructions are obtained to optimize the communication environment. The data transmission parameters are adjusted according to the instructions to implement the measures, the adjusted network throughput index is obtained, it is determined whether the throughput meets the requirements of the production instructions, and the optimization confirmation result is used for the loop feedback of link health assessment.

2. The IoT industrial data processing system and method for multi-device access as described in claim 1, characterized in that, The process involves acquiring real-time network signal strength and latency metrics from a multi-device access industrial IoT environment, determining the communication link health level through comparison with preset thresholds, and obtaining a link health assessment result for subsequent processing, including: Continuously collect real-time network signal strength and latency metrics for each communication link in a multi-device industrial IoT environment; Based on the collected signal strength and delay indicators, they are compared with the corresponding preset thresholds; If the signal strength is lower than a preset threshold, the link signal strength is marked as abnormal. If the latency index exceeds the preset threshold, the link latency is marked as abnormal. The health level of the communication link is determined by the combination of anomaly markers and classified into healthy, slightly abnormal, and severely abnormal. The isolated forest algorithm is used to detect outliers in the signal strength and delay index sequences acquired in multiple time series, correcting misjudgments caused by single threshold comparison. Update the communication link health assessment results based on the corrected anomalies and health levels; The random forest algorithm is used to analyze the correlation between link health assessment results and historical equipment operating status, predict the short-term trend of link health level changes, and generate link health prediction results.

3. The IoT industrial data processing system and method for multi-device access as described in claim 1, characterized in that, The process involves determining network status changes based on the link health assessment results. If the network status change exceeds a preset threshold, a classification algorithm is used to analyze historical data to determine communication quality fluctuation patterns. The resulting fluctuation pattern identification results are then used for strategy formulation, including: Real-time evaluation results data are obtained through the link health monitoring system, and the data is preliminarily processed to obtain the current performance of the network status. Based on the current state of the network, the magnitude of the state change is calculated. If the magnitude exceeds a preset threshold, a subsequent analysis process is triggered to determine the preliminary range of the abnormal fluctuation. To determine the initial range of abnormal fluctuations, corresponding historical data is retrieved from the repository, and the data is classified using a support vector machine algorithm to obtain the fluctuation pattern of communication quality. Based on the fluctuation pattern of communication quality, feature data of key time periods are extracted, and it is determined whether the feature data conforms to the known anomaly type. If it does, a corresponding pattern recognition label is generated. By identifying patterns and labels, matching them with a pre-established strategy library, obtaining adjustment schemes related to the labels, and determining the specific direction of strategy adjustment; After obtaining the specific direction of the strategy adjustment, the corresponding control command is generated and sent to the network management system to complete the optimization of communication quality. By optimizing the processed network status data, we continuously monitor the performance of link health, update the evaluation results database, and maintain the accuracy of real-time data.

4. The IoT industrial data processing system and method for multi-device access as described in claim 1, characterized in that, The process of obtaining priority labels for key data based on the fluctuation pattern recognition results, analyzing the risk level of data loss and delay using a decision tree algorithm, determining whether to activate an alternative transmission path, and obtaining a path activation decision to maintain data reliability includes: Obtain the fluctuation pattern recognition results; Risk assessment is obtained by analyzing the volatility pattern identification results using decision tree analysis; If the risk assessment indicates a risk of loss, then the backup path will be activated. Risk assessment is categorized using priority labeling. Determine path activation decisions based on label classification; The path activation decision is used to obtain transmission control commands. Data reliability maintenance is performed by transmitting control commands.

5. The IoT industrial data processing system and method for multi-device access according to claim 1, characterized in that, If the path activation decision indicates a switch, then available bandwidth information is obtained from the backup path, bandwidth resources are grouped using a clustering algorithm, the optimal transmission channel is determined, and a channel allocation scheme is obtained to connect the production instruction execution process, including: If the path activation decision indicates a switch, obtain available bandwidth information from the backup path; Calculate the current load ratio of each backup path based on available bandwidth information; If the load ratio is lower than the preset threshold, the corresponding backup path is marked as a candidate path; Obtain the bandwidth resource sequence from the candidate paths; Clustering algorithms are used to group the bandwidth resource sequences. The clustering results were used to identify the group with the highest bandwidth stability within the group. Select the transmission channel with the largest peak bandwidth from the highest stability group as the optimal channel; The optimal channel allocation scheme is used to connect the production instruction execution process.

6. The IoT industrial data processing system and method for multi-device access as described in claim 1, characterized in that, The process of transmitting data packets collected by sensors through the channel allocation scheme, obtaining a packet loss rate index during transmission, determining whether the packet loss rate is lower than a preset threshold, and obtaining a transmission reliability verification result for dynamic strategy adjustment includes: Sensor data packets are transmitted using a channel allocation scheme, and the timestamp of each data packet's transmission and receipt confirmation is recorded. Calculate the data packet transmission delay from the sending timestamp and the receiving acknowledgment timestamp, and count the number of data packets that have not received an acknowledgment; The packet loss rate is calculated based on the number of unacknowledged data packets and the total number of data packets sent. Determine whether the packet loss rate is lower than a preset threshold. If it is lower than the preset threshold, the transmission reliability result is determined to be high reliability. Based on the transmission reliability results, collect the channel occupancy rate and interference intensity under the current channel allocation scheme; A random forest model is trained using channel occupancy and interference intensity to obtain channel quality prediction scores; All available channels are sorted according to their predicted channel quality scores, and the channel allocation scheme is adjusted to prioritize channels with high scores.

7. The IoT industrial data processing system and method for multi-device access according to claim 1, characterized in that, The real-time monitoring mechanism, updated based on the transmission reliability verification result, if a continuous change in network status is detected, retrieves interference source data from the device logs, determines interference mitigation measures, and obtains measure execution instructions to optimize the communication environment, including: The current transmission reliability verification results are obtained through a real-time monitoring mechanism; If the transmission reliability verification result is lower than the preset threshold, then determine whether the network status is continuously changing; When the network status continues to change, extract interference source data from the device logs; Based on the extracted interference source data, determine whether the interference type is co-channel interference or multipath interference; If it is determined to be co-channel interference, then frequency switching mitigation measures will be adopted. If it is determined to be multipath interference, then the mitigation measure of adjusting the equalizer parameters is adopted. Based on the determined mitigation measures, corresponding implementation instructions are generated and sent to the communication module.

8. The IoT industrial data processing system and method for multi-device access according to claim 1, characterized in that, The process of adjusting data transmission parameters according to the instructions for implementing the measures, obtaining the adjusted network throughput index, determining whether the throughput meets the requirements for executing production instructions, and obtaining optimization confirmation results for link health assessment includes a cyclical feedback mechanism: For the data transmission process, the current network throughput data is obtained through a pre-established monitoring module to determine the initial baseline values ​​of the transmission parameters; Based on the initial transmission parameter baseline values, an automated tool is used to simulate and calculate the parameter adjustment scheme to obtain the adjusted parameter configuration scheme; For the adjusted parameter configuration scheme, the scheme is loaded into the simulation environment through network transmission, and the changes in throughput data after the simulation run are obtained; Based on the changes in throughput data after the simulation, if the throughput data is detected to be lower than the preset execution standard, the parameter readjustment process is triggered to determine a new parameter configuration scheme. For the new parameter configuration scheme, it is deployed in a real network environment to obtain real-time throughput data and determine whether it meets the execution standards of production instructions; Based on the judgment results of real-time throughput data, if the execution criteria are met, the parameter configuration scheme is recorded in the link health assessment database to obtain the preliminary assessment results of the link health status. Based on the preliminary assessment results of the link health status, the assessment loop data is transmitted to the parameter adjustment module through a feedback mechanism to determine the optimization direction for the next data transmission.

Images