Data transmission link quality intelligent monitoring optimization system based on ai technology

Abstract

The application belongs to the technical field of data transmission supervision, and specifically relates to an intelligent data transmission link quality monitoring and optimization system based on AI technology, which comprises a link feature sensing and collecting module, an AI adaptive noise reduction preprocessing module, a link quality multi-dimensional evaluation module, a link degradation root cause tracing and positioning module, a transmission channel intelligent adaptive scheduling module, and a link self-healing strategy generation and execution module; the application collects link core features and completes AI adaptive noise reduction, quantitatively evaluates link quality and classifies the link quality through an AI model, locates degradation root causes relying on AI inference, filters optimal channels according to link states and root causes to perform differentiated scheduling, implements online lightweight or offline deep self-healing for risk / low-quality links and checks optimization effects, significantly improves data transmission stability and resource utilization, reduces operation and maintenance costs, and is conducive to realizing full-process unmanned intelligent management and control of data transmission links.

Description

Technical Field

[0001] This invention relates to the field of data transmission monitoring technology, specifically to an AI-based intelligent monitoring and optimization system for data transmission link quality. Background Technology

[0002] Against the backdrop of rapid development in digitalization and networking, the scale of data transmission and business scenarios continue to expand. The transmission rate, stability, and reliability of data transmission links directly determine the efficiency of data interaction and the security of business operations. Comprehensive monitoring, accurate evaluation, intelligent scheduling, and autonomous self-healing of link quality have become core requirements for ensuring the efficient and stable operation of data transmission networks.

[0003] Current traditional data transmission link quality control still relies on manual monitoring, fixed threshold alarms, and passive repairs. It has not integrated AI technology to build a fully intelligent control system, making it difficult to adapt to complex and ever-changing link control needs. It generally suffers from the following core defects: First, the link quality assessment method is crude, relying on simple calculation of a single indicator and failing to build a multi-dimensional comprehensive assessment model through AI algorithms. It cannot achieve accurate quality quantification and scientific classification, and the root cause of degradation cannot be traced intelligently. It can only trigger basic alarms and it is difficult to locate the root cause of degradation through AI feature association reasoning. Fault diagnosis relies too much on manual labor. Secondly, channel scheduling lacks intelligent adaptation, adopts fixed switching or manual intervention mode, does not combine dual-dimensional indicators to select the optimal channel, resource allocation is unreasonable and utilization is low, and link self-healing is undifferentiated and lacks verification. It does not implement hierarchical self-healing based on link status and root causes of degradation and does not automatically verify the self-healing effect, which cannot guarantee the effectiveness of self-healing and makes it difficult to achieve unmanned, intelligent and full-process control of data transmission links.

[0004] Therefore, developing a data transmission link quality intelligent monitoring and optimization system capable of achieving multi-dimensional AI assessment of link quality, intelligent tracing of the root causes of degradation, intelligent adaptation and scheduling of transmission channels, and differentiated link self-healing execution has become an urgent technical problem to be solved in the field of data transmission supervision. Summary of the Invention

[0005] The purpose of this invention is to provide an intelligent monitoring and optimization system for data transmission link quality based on AI technology, so as to solve the technical defects mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: an AI-based intelligent monitoring and optimization system for data transmission link quality, comprising a link feature perception and acquisition module, an AI adaptive noise reduction preprocessing module, a multi-dimensional link quality evaluation module, a link degradation root cause tracing and location module, a transmission channel intelligent adaptation and scheduling module, and a link self-healing strategy generation and execution module. The link feature perception and acquisition module collects multi-type core feature data of the data transmission link in real time. The AI ​​adaptive noise reduction preprocessing module uses AI adaptive noise reduction algorithm to perform differentiated noise reduction processing and output stable and accurate feature data. The link quality multi-dimensional assessment module combines AI algorithm to build a multi-dimensional assessment model and performs accurate quantitative assessment of link quality based on the preprocessed feature data. The link degradation root cause tracing and positioning module accurately locates the core causes of link quality degradation based on the link quality classification results and normalized data of each indicator through AI feature association reasoning and root cause contribution calculation. The intelligent adaptation and scheduling module for transmission channels calculates the channel adaptation matching degree based on the link quality classification and the root cause of degradation, combined with the idle channel resource pool, based on the dual dimensions of link quality fit and the historical operational stability of backup channels, selects the optimal adapted channel for use, and performs traffic splitting and scheduling; the link self-healing strategy generation and execution module generates and executes an adaptive self-healing adjustment strategy for links with different quality levels and root causes of degradation, combined with the link service carrying status.

[0007] Furthermore, the specific process by which the multi-dimensional link quality assessment module performs precise quantitative assessment of link quality is as follows: The system receives feature data transmitted by the AI ​​adaptive noise reduction preprocessing module, extracts six core evaluation indicators of the data transmission link: transmission rate V, bit error rate E, transmission delay T, signal jitter J, signal strength S, and link load L, and normalizes each indicator so that the value of each indicator is within the range [0,1], thereby obtaining the normalized values ​​of the six indicators. A multi-dimensional evaluation model is constructed by combining AI algorithms, and the comprehensive link quality evaluation value Q is obtained by weighted summation. After calculating the comprehensive link quality evaluation value Q, the data transmission link is classified according to the link quality evaluation and grading standard. After the evaluation is completed, the evaluation information, including the Q value, the normalization results of each indicator, and the link quality classification results, will be synchronously transmitted to the link degradation root cause tracing and location module, the transmission channel intelligent adaptation and scheduling module, and the link self-healing strategy generation and execution module.

[0008] Furthermore, the link quality assessment and grading standards are as follows: If Q ≥ 0.8, then the corresponding data transmission link is judged to be a high-quality link; If 0.6 ≤ Q < 0.8, then the corresponding data transmission link is judged to be a good link; If 0.4 ≤ Q < 0.6, then the corresponding data transmission link is judged to be a risky link; If Q < 0.4, the corresponding data transmission link is judged to be a poor quality warning link.

[0009] Furthermore, after receiving the evaluation information transmitted by the multi-dimensional link quality evaluation module, the link degradation root cause tracing and localization module retrieves the system's historical fault root cause sample library, inputs the current normalized values ​​of various indicators and preset indicator weights into the AI ​​correlation inference model, and calculates the contribution of each indicator to the link quality degradation. The contribution values ​​of the six indicators are traversed, and the top three indicators in terms of contribution are identified as degradation-related indicators. Then, the root cause classification is completed by combining the indicator types. After the root cause is located, the source location information, including the degradation root cause type and the core related indicator data, is transmitted to the transmission channel intelligent adaptation and scheduling module and the link self-healing strategy generation and execution module.

[0010] Furthermore, the specific process by which the intelligent adaptation and scheduling module for selecting the optimal adaptation and allocation channel is as follows: The real-time comprehensive quality assessment value Qs of each backup channel is retrieved from the multi-dimensional link quality assessment module. If Qs≥0.6, the corresponding backup channel is marked as a qualified backup channel, and an idle channel resource list is constructed based on the qualified backup channels. The operational stability coefficient S of each qualified backup channel is obtained through analysis and calculation. Combined with the comprehensive evaluation value Qc of the current data transmission link quality, the comprehensive compatibility degree between the current link and each qualified backup channel is calculated using the channel compatibility matching degree formula. The qualified backup channel with the largest channel compatibility matching degree value is selected and marked as the optimal adapted backup channel for the data transmission link.

[0011] Furthermore, after the intelligent adaptation and scheduling module for transmission channels completes the optimal adaptation and allocation channel selection, it executes a differentiated scheduling strategy based on link quality, root causes of degradation, and service transmission priorities. After scheduling is completed, scheduling monitoring information, including channel switching records, link quality comparison data before and after scheduling, and service diversion ratios, is transmitted to the link self-healing strategy generation and execution module. The scheduling strategy is as follows: If the current data transmission link is a risky link and the root cause is channel congestion, low-priority service traffic will be scheduled to the optimally adapted allocated channel, while high-priority services will be retained for transmission on the original link. If the current data transmission link is a poor quality warning link, all service traffic will be directly migrated to the optimally adapted allocated channel, and the service carrying authority of the poor quality warning link will be temporarily blocked. If the current data transmission link is a high-quality or good link, the original channel allocation will be maintained, and only real-time monitoring will be performed without any channel scheduling or switching operations.

[0012] Furthermore, the analysis and calculation method for the operational stability coefficient S is as follows: Retrieve the quality assessment records and fault event records of each qualified backup channel for the past thirty days from the system's historical operation database. Calculate the cumulative normal operation time Tnorm of the corresponding qualified backup channels with a comprehensive quality assessment value greater than or equal to 0.6 in the past thirty days. Calculate the ratio of the cumulative normal operation time Tnorm to the total operation time Ttotal in the past thirty days to obtain the percentage of normal operation time. Then, based on the channel's historical degradation and fault conditions, a corresponding fault penalty coefficient Kfault is matched. The proportion of normal operating time is multiplied by the fault penalty coefficient Kfault to obtain the corresponding operational stability coefficient S of the qualified backup channel.

[0013] Furthermore, the link self-healing strategy generation and execution module performs differentiated self-healing operations based on the link classification results, the type of root cause of degradation, and whether the link carries services: For risky links that are still carrying data transmission, online lightweight self-healing processing is carried out. If the root cause of the degradation is external electromagnetic interference, the link signal verification strength is slightly increased and the signal filtering parameters are optimized. If the root cause of the degradation is high link load, the data transmission time slot is dynamically adjusted and the data packet distribution logic is optimized. The existing data transmission is not interrupted throughout the process, and the link transmission quality is gradually improved by fine-tuning the parameters. For poor-quality early warning links that have completed business traffic migration and are in an idle state, an offline deep self-healing process is initiated, which sequentially performs self-check of transmission hardware port status, reset of transmission protocol parameters, calibration of signal transmit and receive power, and reconstruction of channel filtering parameters to comprehensively repair abnormal link transmission issues. For links in good or excellent condition, no self-healing operation is triggered; only routine link quality monitoring is maintained.

[0014] Furthermore, after the self-healing is completed, the link self-healing strategy generation and execution module calculates the improvement in link quality before and after the self-healing optimization, and calculates the quality improvement difference ΔQ; if ΔQ≤0, the current self-healing strategy and scheduling scheme are determined to be invalid, and self-healing alarm information is output.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In this invention, an AI model is used to accurately quantify and classify link quality and quickly locate the root cause of degradation. The optimal channel is selected based on the dual dimensions of channel adaptability and operational stability, and differentiated traffic scheduling is performed. This significantly improves the accuracy of link quality assessment and channel resource utilization, ensures stable and efficient data transmission, and facilitates the realization of unmanned intelligent management and control of the entire data transmission link process.

[0016] 2. In this invention, through the differentiated link self-healing function, risky links carrying services are self-healed online in a lightweight manner, and idle and inferior links are self-healed offline in a deep manner. The self-healing effect is automatically verified and abnormal alarms are triggered, realizing the fully automated repair of link failures. This greatly reduces manual operation and maintenance intervention and lowers operation and maintenance costs, further enhancing the stability and self-healing reliability of data transmission links. Attached Figure Description

[0017] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a schematic diagram of the overall system structure of the present invention; Figure 2 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: Refer to Figure 1-2 As shown, the intelligent monitoring and optimization system for data transmission link quality based on AI technology proposed in this invention includes a link feature perception and acquisition module, an AI adaptive noise reduction preprocessing module, a multi-dimensional link quality evaluation module, a link degradation root cause tracing and location module, a transmission channel intelligent adaptation and scheduling module, and a link self-healing strategy generation and execution module. The link feature perception and acquisition module collects multiple types of core feature data of the data transmission link in real time, realizing data collection in all scenarios without blind spots. It can dynamically adjust the acquisition frequency according to the link load, eliminate abnormal data, and provide high-quality raw data for subsequent preprocessing and evaluation, which is different from the limitations of existing technologies that collect single indicators and collect data at fixed frequencies.

[0020] Specifically, the link feature perception and acquisition module has a built-in multi-channel adaptive acquisition unit that synchronously acquires six core feature data of the data transmission link: transmission rate, bit error rate, transmission delay, signal jitter, signal strength, and link load. Before acquisition, the link load value is obtained in real time through the built-in link load detection unit. If the link load is ≥80%, the acquisition frequency is adjusted to 100 times / second to ensure the real-time data acquisition in high-load scenarios; if the link load is <80%, the acquisition frequency is adjusted to 20 times / second to reduce system resource consumption in low-load scenarios.

[0021] Furthermore, during the acquisition process, the module performs preliminary verification on the acquired raw data, removing invalid data with negative transmission rates, bit error rates exceeding reasonable thresholds, or abnormal latency jumps. The six types of verified feature data are then encapsulated into standardized data frames. Each data frame contains an acquisition timestamp, link identifier, and feature data fields. After encapsulation, the data is transmitted in real time to the AI ​​adaptive noise reduction preprocessing module through an internal data interface, providing raw data support for subsequent noise reduction processing.

[0022] The AI ​​adaptive noise reduction preprocessing module receives the raw data transmitted by the link feature perception acquisition module, identifies random noise and system noise in the data, and uses the AI ​​adaptive noise reduction algorithm to perform differentiated noise reduction processing to eliminate data deviations caused by electromagnetic interference and transmission loss, and outputs stable and accurate feature data, thus solving the problems of weak adaptability and poor noise reduction effect of existing technologies with fixed noise reduction rules.

[0023] Specifically, after receiving the standardized data frame transmitted by the link feature perception and acquisition module, the AI ​​adaptive noise reduction preprocessing module first splits the acquisition timestamp, link identifier, and 6 types of feature data through the data parsing unit, and then inputs the feature data into the trained AI noise recognition model. The model distinguishes the noise type based on the data fluctuation amplitude and the duration of abnormality. If the data changes randomly in a short period of time, it is judged as random noise. If the data continuously deviates from the standard value, it is judged as system noise.

[0024] To address random noise, a moving average filtering algorithm is used, calculating the average value of data from five consecutive acquisition periods and replacing the noisy data. For system noise, a Transformer-based adaptive denoising model is employed, retrieving historical denoising parameters from the same scene for fitting and correction to eliminate persistent noise interference. After denoising, the processed data undergoes secondary verification to ensure the feature data is within a reasonable value range. Once verification is successful, the denoised feature data, link identifier, and acquisition timestamp are synchronously transmitted to the multi-dimensional link quality evaluation module, providing accurate data for link quality assessment.

[0025] The multi-dimensional link quality assessment module combines AI algorithms to build a multi-dimensional assessment model. Based on preprocessed feature data, it performs accurate quantitative assessment of link quality and outputs assessment results that can be directly used for subsequent optimization, solving the problems of single assessment dimensions and low assessment accuracy in existing technologies.

[0026] Specifically, the multi-dimensional link quality evaluation module first receives the feature data transmitted by the AI ​​adaptive noise reduction preprocessing module, and extracts six core evaluation indicators of the data transmission link: transmission rate V, bit error rate E, transmission delay T, signal jitter J, signal strength S, and link load L. In order to eliminate the influence of the difference in the dimensions of different indicators on the evaluation results, each indicator is normalized (minimum-maximum normalization) so that the value of each indicator is within the range [0,1], and the normalized values ​​of the six indicators Y1, Y2, Y3, Y4, Y5 and Y6 are obtained respectively. A multi-dimensional evaluation model is constructed by combining AI algorithms. The comprehensive link quality evaluation value Q is obtained by weighted summation, as shown in the following formula: Q=W1×Y1+W2×Y2+W3×Y3+W4×Y4+W5×Y5+W6×Y6; Q: Comprehensive link quality assessment value, ranging from [0,1]; Among them, W1, W2, W3, W4, W5, W6: preset indicator weight values, and W1+W2+W3+W4+W5+W6=1; After calculating the comprehensive link quality assessment value Q, the data transmission link is classified according to the link quality assessment grading standard. The grading strategy can be referred to as follows: If Q ≥ 0.8, then the corresponding data transmission link is judged to be a high-quality link; If 0.6 ≤ Q < 0.8, then the corresponding data transmission link is judged to be a good link; If 0.4 ≤ Q < 0.6, then the corresponding data transmission link is judged to be a risky link; If Q < 0.4, the corresponding data transmission link is judged to be a poor quality warning link.

[0027] Furthermore, after the evaluation is completed, the evaluation information, including the Q value, the normalization results of each indicator, and the link quality classification results, will be synchronously transmitted to the link degradation root cause tracing and localization module and the transmission channel intelligent adaptation and scheduling module, providing core basis for subsequent link problem analysis.

[0028] The link degradation root cause tracing and localization module, based on the link quality grading results and normalized data of various indicators, accurately locates the core causes of link quality degradation through AI feature association reasoning and root cause contribution calculation. It distinguishes between link hardware problems, external interference problems, and load congestion problems, solving the technical pain point of existing technologies that can only identify quality degradation but cannot locate the root cause.

[0029] Specifically, after receiving the evaluation information transmitted by the multi-dimensional link quality evaluation module, the link degradation root cause tracing and localization module first retrieves the system's historical fault root cause sample database, inputs the current normalized values ​​of each indicator and the preset indicator weights into the AI ​​correlation inference model, and calculates the contribution of each indicator to the link quality degradation. The root cause contribution calculation formula is as follows: G = Wi × |Yio - Yi|; Among them, G: contribution of single indicator degradation. It should be noted that the larger the G value, the more likely that the indicator is the core cause of the link quality degradation. Wi: The preset indicator weight value of the corresponding indicator, which can be directly transmitted and obtained from the multi-dimensional link quality evaluation module; Yio: The standard normalized value of the corresponding indicator under excellent conditions, which is pre-set and stored; Yi: The current normalized value of the corresponding indicator can be directly obtained from the multi-dimensional link quality evaluation module. Iterate through the contribution values ​​of the six indicators, identify the top three indicators as deterioration-related indicators, and then combine the indicator types to complete the root cause classification. See below for reference: If the contribution of bit error rate and signal strength indicators is high, it is determined to be external electromagnetic interference or link transmission loss; if the contribution of transmission delay and link load indicators is high, it is determined to be channel congestion or service traffic overload; if the signal jitter indicator is continuously abnormal, it is determined to be transmission hardware performance degradation. Furthermore, after the root cause is located, the source location information, including the type of deterioration root cause and the core related indicator data, is transmitted to the transmission channel intelligent adaptation and scheduling module.

[0030] The intelligent adaptation and scheduling module for transmission channels calculates the channel adaptation matching degree based on the link quality classification and the root cause of degradation, combined with the idle channel resource pool, based on the dual dimensions of link quality fit and the historical operational stability of backup channels. It then selects the optimal adapted channel and performs traffic splitting and scheduling to avoid inferior links from continuously carrying critical services, thereby improving the resource utilization and stability of the overall transmission network.

[0031] Specifically, after receiving the link classification, root cause and index data transmitted by the link degradation root cause tracing and positioning module, the transmission channel intelligent adaptation and scheduling module retrieves the real-time comprehensive quality evaluation value Qs of each backup channel from the link quality multi-dimensional evaluation module. If Qs≥0.6, the corresponding backup channel is marked as a qualified backup channel, and an idle channel resource list is constructed based on the qualified backup channels. The operational stability coefficient S of each qualified backup channel was obtained through analysis and calculation: Retrieve the quality assessment records and fault event records of each qualified backup channel for the past thirty days from the system's historical operation database. Calculate the cumulative normal operation time Tnorm of the corresponding qualified backup channels with a comprehensive quality assessment value greater than or equal to 0.6 in the past thirty days. Calculate the ratio of the cumulative normal operation time Tnorm to the total operation time Ttotal in the past thirty days to obtain the percentage of normal operation time. Then, based on the channel's historical degradation and fault conditions, a corresponding fault penalty coefficient Kfault is matched. The operational stability coefficient S of the corresponding qualified backup channel is obtained by multiplying the percentage of normal operating time by the fault penalty coefficient Kfault; that is... ; It should be noted that if there are no risk or quality warnings or faults in the past 30 days, Kfault = 1.0; if there are no more than 2 risk warnings (0.4 ≤ Q < 0.6) in the past 30 days, Kfault = 0.9; if there is 1 or more quality warnings (Q < 0.4) in the past 30 days, Kfault = 0.7; if there has been a transmission interruption or link disconnection fault in the past 30 days, Kfault = 0.5.

[0032] After calculating the operational stability coefficient S of the corresponding qualified backup channel, and combining it with the comprehensive link quality evaluation value Qc of the current data transmission link, the comprehensive compatibility degree between the current link and each qualified backup channel is calculated using the channel adaptation matching degree formula, as follows: R = e1 × (1 - |Qs - Qc|) + e2 × S; Where R: channel adaptation matching degree, with a value of [0,1]. The higher the value, the better the backup channel matches the current link and the more reliable the operation. e1, e2: Preset weighting coefficients, with the default values ​​of e1=0.7 and e2=0.3, and their sum is always 1; Select the qualified backup channel with the highest channel fit matching degree (the higher the matching degree value, the higher the degree of fit between the corresponding qualified backup channel and the current data transmission link, and the more reliable its own historical operating status), and mark it as the optimal matching backup channel for the data transmission link.

[0033] Furthermore, after selecting the optimal allocated channel, a differentiated scheduling strategy is implemented based on link quality, root causes of degradation, and service transmission priority; the scheduling strategy is as follows: If the current data transmission link is a risky link and the root cause is channel congestion, low-priority service traffic will be scheduled to the optimally adapted allocated channel, while high-priority services will be retained for transmission on the original link. If the current data transmission link is a poor quality warning link, all service traffic will be directly migrated to the optimally adapted allocated channel, and the service carrying authority of the poor quality warning link will be temporarily blocked to prevent transmission interruption. If the current data transmission link is a high-quality or good link, the original channel allocation will be maintained, and only real-time monitoring will be performed without any channel scheduling or switching operations.

[0034] Example 2: Refer to Figure 1-2 As shown, the difference between this embodiment and Embodiment 1 is that it also includes a link self-healing strategy generation and execution module. After the intelligent adaptation and scheduling module of the transmission channel completes the scheduling, it transmits the scheduling monitoring information, including channel switching records, link quality comparison data before and after scheduling, and service diversion ratio, to the link self-healing strategy generation and execution module. The multi-dimensional evaluation module of link quality and the root cause tracing and location module of link degradation transmit the quality evaluation information and the tracing and location information to the link self-healing strategy generation and execution module. The link self-healing strategy generation and execution module generates and executes adaptive self-healing adjustment strategies for links with different quality levels and root causes of degradation, taking into account the link service carrying status. This reduces manual intervention and automatically restores link transmission performance, adapting to complex and ever-changing transmission scenarios and significantly reducing manual operation and maintenance costs.

[0035] Specifically, the link self-healing strategy generation and execution module performs differentiated self-healing operations based on the link classification results, the type of root cause of degradation, and whether the link carries services: For risky links that are still carrying data transmission, online lightweight self-healing processing is carried out. For example, if the root cause of degradation is external electromagnetic interference, the link signal verification strength is slightly increased and the signal filtering parameters are optimized; if the root cause of degradation is high link load, the data transmission time slot is dynamically adjusted and the data packet distribution logic is optimized. The existing data transmission is not interrupted throughout the process, and the link transmission quality is gradually improved by fine-tuning the parameters. For poor-quality early warning links that have completed business traffic migration and are in an idle state, an offline deep self-healing process is initiated, which sequentially performs self-check of transmission hardware port status, reset of transmission protocol parameters, calibration of signal transmit and receive power, and reconstruction of channel filtering parameters to comprehensively repair abnormal link transmission issues. For links in good or excellent condition, no self-healing operation is triggered; only routine link quality monitoring is maintained.

[0036] Furthermore, after the self-healing process is completed, the link self-healing strategy generation and execution module calculates the improvement in link quality before and after self-healing optimization. The calculation formula is as follows: ; Wherein, ΔQ: the difference in quality improvement; Q t : Comprehensive evaluation value of link quality after self-healing execution; Q o : Initial link quality comprehensive assessment value before self-healing execution; If ΔQ≤0, the current self-healing strategy and scheduling scheme are deemed invalid, and a self-healing alarm is output so that manual intervention can be carried out to make corresponding improvement and optimization measures, further ensuring the stable and efficient operation of the link.

[0037] The working principle of this invention is as follows: In use, six core feature data of the transmission link are collected in real time. After AI adaptive noise reduction processing, the link quality is quantitatively evaluated and graded through an AI multi-dimensional model. The root cause of link degradation is located based on AI feature correlation reasoning. The optimal backup channel is selected based on quality fit and channel stability, and differentiated traffic scheduling is performed. Then, online lightweight or offline deep self-healing is carried out on risky / inferior links respectively, and the optimization effect is verified. This invention solves the problems of single dimension of traditional link monitoring and evaluation, difficulty in locating the root cause of degradation, and crude channel scheduling and self-healing methods. It significantly improves the accuracy of link quality assessment, the efficiency of root cause location, and the utilization rate of channel resources, reduces manual intervention and lowers operation and maintenance costs, ensures stable and efficient data transmission, and is conducive to realizing unmanned intelligent management and control of the entire data transmission link process.

[0038] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, enabling those skilled in the art to better understand and utilize it. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A data transmission link quality intelligent monitoring and optimization system based on AI technology, characterized in that: It includes a link feature perception and acquisition module, an AI adaptive noise reduction preprocessing module, a multi-dimensional link quality assessment module, a link degradation root cause tracing and location module, a transmission channel intelligent adaptation and scheduling module, and a link self-healing strategy generation and execution module. The link feature perception and acquisition module collects multi-type core feature data of the data transmission link in real time. The AI ​​adaptive noise reduction preprocessing module uses AI adaptive noise reduction algorithm to perform differentiated noise reduction processing and output stable and accurate feature data. The link quality multi-dimensional assessment module combines AI algorithm to build a multi-dimensional assessment model, and performs accurate quantitative assessment of link quality based on the preprocessed feature data to generate link quality grading results. The link degradation root cause tracing and positioning module accurately locates the core causes of link quality degradation based on the link quality grading results and normalized data of each indicator through AI feature association reasoning and root cause contribution calculation. The intelligent adaptation and scheduling module of the transmission channel calculates the channel adaptation matching degree based on the link quality classification results and the root cause results of degradation, combined with the idle channel resource pool, based on the link quality matching degree and the historical operation stability of the backup channel, selects the optimal adapted and used channel and performs traffic splitting and scheduling. The link quality matching degree is the difference between the real-time comprehensive quality evaluation value of each backup channel and the comprehensive link quality evaluation value of the current data transmission link. The link self-healing strategy generation and execution module generates and executes adaptive self-healing adjustment strategies for links with different quality levels and root causes of degradation, taking into account the link service carrying status. The specific process for accurately quantifying and evaluating link quality is as follows: Six core evaluation indicators for data transmission links are extracted: transmission rate V, bit error rate E, transmission delay T, signal jitter J, signal strength S, and link load L. Each indicator is then normalized. A multi-dimensional evaluation model is constructed using AI algorithms, and the comprehensive link quality evaluation value Q is obtained through weighted summation. After calculating the comprehensive link quality assessment value Q, the data transmission link is graded based on the link quality assessment grading standard. After the evaluation is completed, the comprehensive link quality evaluation value Q and the quality classification results will be synchronously transmitted to the link degradation root cause tracing and location module, the transmission channel intelligent adaptation and scheduling module, and the link self-healing strategy generation and execution module. The link quality assessment and grading standards are as follows: If Q ≥ 0.8, the corresponding data transmission link is judged as a high-quality link; if 0.6 ≤ Q < 0.8, the corresponding data transmission link is judged as a good link; if 0.4 ≤ Q < 0.6, the corresponding data transmission link is judged as a risky link. If Q < 0.4, the corresponding data transmission link is judged to be a poor quality warning link.

2. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 1, characterized in that, The link degradation root cause tracing and localization module inputs the normalized values ​​of various indicators and the preset indicator weights into the AI ​​correlation inference model to calculate the contribution of each indicator to the link quality degradation. The contribution values ​​of the six indicators are traversed, and the top three indicators in terms of contribution are identified as degradation-related indicators. Then, the root cause classification is completed by combining the indicator types. After the root cause is located, the source location information is transmitted to the transmission channel intelligent adaptation and scheduling module and the link self-healing strategy generation and execution module.

3. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 2, characterized in that, The specific process of the transmission channel intelligent adaptation and scheduling module in selecting the optimal adapted channel is as follows: Based on the qualified backup channels, a list of idle channel resources is constructed. The operational stability coefficient S of each qualified backup channel is obtained through analysis and calculation. Combined with the comprehensive evaluation value Qc of the current data transmission link quality, the comprehensive adaptation degree between the current link and each qualified backup channel is calculated. The qualified backup channel with the largest channel adaptation matching degree value is selected and marked as the optimal adapted channel.

4. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 3, characterized in that, After the intelligent adaptation and scheduling module for transmission channels completes the optimal adaptation and allocation channel selection, it executes a differentiated scheduling strategy. Upon completion of scheduling, it transmits the scheduling monitoring information to the link self-healing strategy generation and execution module. The scheduling strategy is as follows: If the current data transmission link is a risky link and the root cause is channel congestion, low-priority service traffic will be scheduled to the optimally adapted allocated channel, while high-priority services will be retained for transmission on the original link. If the current data transmission link is a poor quality warning link, all service traffic will be directly migrated to the optimally adapted allocated channel, and the service carrying authority of the poor quality warning link will be temporarily blocked. If the current data transmission link is a high-quality or good link, the original channel allocation will be maintained, and only real-time monitoring will be performed without any channel scheduling or switching operations.

5. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 4, characterized in that, The method for analyzing and calculating the operational stability coefficient S is as follows: The cumulative normal operating time of the corresponding qualified backup channels with a comprehensive quality assessment value greater than or equal to 0.6 within the past thirty days is statistically analyzed. The ratio of the cumulative normal operating time to the total operating time within the past thirty days is calculated to obtain the normal operating time percentage. The normal operating time percentage is multiplied by the fault penalty coefficient to obtain the operating stability coefficient S of the corresponding qualified backup channel.

6. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 5, characterized in that, The link self-healing strategy generation and execution module performs differentiated self-healing operations based on the link classification results, the type of root cause of degradation, and whether the link carries services: For risky links that are still carrying data transmission, implement online lightweight self-healing processing; For substandard early warning links that have completed business traffic migration and are in an idle state, initiate an offline deep self-healing process; For links that are in good or excellent condition, no self-healing operation is triggered.

7. The intelligent monitoring and optimization system for data transmission link quality based on AI technology according to claim 6, characterized in that, After the self-healing process is completed, the link self-healing strategy generation and execution module calculates the improvement in link quality before and after the self-healing optimization, and calculates the quality improvement difference ΔQ. If ΔQ≤0, the current self-healing strategy and scheduling scheme are determined to be invalid, and self-healing alarm information is output.

Images