Intelligent diagnosis method and system for fiber communication link fault based on feature fusion

Abstract

The application provides a kind of optical fiber communication link fault intelligent diagnosis method and system based on feature fusion, belongs to communication fault diagnosis technical field, to adopt a kind of predictive dynamic sampling and sampling density adaptive fusion combined diagnostic strategy, can perceive risk in advance before fault occurs, and dynamically adjust monitoring and fusion strategy, improve the timeliness, accuracy and system adaptive ability of diagnosis.The application is through the diagnostic strategy of predictive dynamic sampling and sampling density adaptive fusion, improves the timeliness of optical fiber physical layer fault diagnosis, is no longer passive waiting for optical signal loss fault to occur, but through time series prediction model, early prediction high-risk optical transmission node and forward-lookingly adjust dynamic attention area, realize the early coverage and high-frequency monitoring of potential optical layer fault point, capture the weak omen before optical fiber physical fault occurs, such as optical signal-to-noise ratio degradation and optical attenuation fluctuation, gain response time for optical network physical routing scheduling.

Description

Technical Field

[0001] This invention relates to the field of communication fault diagnosis technology, and in particular to an intelligent diagnosis method and system for optical fiber communication link faults based on feature fusion. Background Technology

[0002] As the infrastructure of modern information society, the stable operation of communication networks is crucial. Fiber optic communication links are the basic units constituting the network; their failure can lead to service interruptions, degraded service quality, and even widespread network paralysis. Therefore, rapid and accurate diagnosis of fiber optic communication link faults is a core task for ensuring network reliability. Intelligent fault diagnosis technology for fiber optic communication links aims to achieve real-time monitoring, type identification, and root cause location of network faults through automated data analysis and model reasoning.

[0003] In existing technologies, fiber optic communication link fault diagnosis typically relies excessively on traffic status data from the network or data link layer. However, in real-world fiber optic transmission scenarios, optical physical degradation, such as fiber micro-bending, connector aging, and the accumulation of spontaneous emission noise from optical amplifiers, is the underlying cause of link transmission interruptions. While existing methods utilize Long Short-Term Memory (LSTM) networks to process time-series performance data and convolutional neural networks to process log text data for heterogeneous data, their data sampling strategies are relatively rigid, typically employing polling at a uniform frequency across the entire network. This approach, lacking dynamic tracking of underlying optical physical characteristics, cannot adapt to dynamic changes in optical channel states and is prone to missing early signals of hidden optical physical faults, such as early power degradation. Furthermore, the feature fusion stage fails to consider the differences in information reliability caused by different sampling frequencies at different optical nodes, leaving room for improvement.

[0004] Existing technologies still have significant shortcomings in practical applications. Data sampling strategies are relatively rigid, typically employing polling at a uniform frequency across the entire network or strengthening monitoring only on a few key fiber optic transmission nodes based on static rule definitions. This approach cannot adapt to dynamic changes in network conditions and is prone to missing early signals of sudden or hidden faults. In the feature fusion stage, existing models often treat all features from all sources equally, failing to consider the differences in information reliability caused by different sampling frequencies. This makes the fusion results susceptible to interference from low-quality data, leaving room for improvement. Summary of the Invention

[0005] This invention provides a method and system for intelligent diagnosis of optical fiber communication link faults based on feature fusion. It employs a diagnostic strategy that combines predictive dynamic sampling with adaptive sampling density fusion, which can detect risks in advance before a fault occurs and dynamically adjust the monitoring and fusion strategy, thereby improving the timeliness, accuracy and adaptability of the diagnosis.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: Firstly, a method for intelligent diagnosis of optical fiber communication link faults based on feature fusion is provided. This method is applied to a fault diagnosis terminal and includes the following steps: The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node are obtained from the optical fiber communication link. The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes are weighted and summed to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node. Based on the fiber optic link failure tendency value, high-risk nodes are selected from all fiber optic transmission nodes as seed nodes, and dynamic interest areas are delineated based on the seed nodes and their proximity connections in the network topology, generating regional attribute data containing the division status of fiber optic transmission nodes. A high-frequency sampling command is sent to the optical fiber transmission nodes in the dynamic interest area to obtain high-frequency sampling data, and a low-frequency sampling command is sent to the optical fiber transmission nodes in the non-interest area to obtain low-frequency sampling data. The high-frequency sampling data and the low-frequency sampling data are then merged to generate a multi-level sampling dataset. The fiber optic link failure tendency value generated by each fiber optic transmission node within a continuous historical time step is extracted to construct a fiber optic link failure tendency value time series. The fiber optic link failure tendency value time series is predicted to obtain the predicted fiber optic link failure tendency value for future time. The boundary of the dynamic attention region is updated according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. Extract the temporal dynamic features from the high-frequency sampled data obtained within the predictive dynamic attention region to generate a first feature vector, and extract the textual semantic features from the low-frequency sampled data obtained within the non-attention region to generate a second feature vector. The first feature vector, the second feature vector, and the numerical sequence representing the change of the identity of the members in the region of the optical fiber transmission node are concatenated to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The high-dimensional fused feature vector is input into the classifier for decoding, and the output includes the fault type and the location of the faulty fiber transmission node. The location of the faulty fiber transmission node is then fed back into the calculation step of the fiber link fault tendency value to dynamically update the weight parameters of the weighted summation.

[0007] Optionally, the step of obtaining the historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and calculating the optical link fault tendency value of each optical fiber transmission node by weighted summation of the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes, specifically includes: Receive the number of faults and total running time reported by each optical fiber transmission node within a specified time window, and calculate the ratio between the two to generate the historical optical attenuation degradation rate. Receive the instantaneous bandwidth occupancy rate and queue length reported by each optical fiber transmission node, and perform a linear combination to generate the current optical channel load; Obtain the fault transfer status of each optical fiber transmission node and its neighboring optical fiber transmission nodes in adjacent time steps, and calculate the crosstalk propagation probability of the adjacent optical nodes. The optical fiber link fault tendency value is generated by multiplying the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes by dynamically adjustable first preset weights, second preset weights, and third preset weights, respectively.

[0008] Optionally, the step of selecting high-risk nodes as seed nodes from all fiber optic transmission nodes based on the fiber optic link failure tendency value, and delineating dynamically concerned regions based on the seed nodes and their proximity connections in the network topology, and generating regional attribute data containing the division status of fiber optic transmission nodes, specifically includes: The fiber optic link failure tendency values ​​of all fiber optic transmission nodes are sorted in descending order, and a specified number of the top-ranked fiber optic transmission nodes are selected as the seed nodes. A dynamic radius proportional to the fiber optic link failure tendency value is calculated for each seed node, and candidate regions are delineated in the network topology based on the dynamic radius. If there are overlapping fiber optic transmission nodes among multiple candidate regions, the candidate regions containing overlapping fiber optic transmission nodes are merged into the same region to generate the dynamic region of interest. Record the list of internal fiber optic transmission nodes, the list of edge fiber optic transmission nodes, and the timestamp of the region formation in the dynamically monitored area, and combine them to generate the region attribute data.

[0009] Optionally, the step of issuing high-frequency sampling commands to fiber optic transmission nodes within the dynamically interested area to obtain high-frequency sampling data, and issuing low-frequency sampling commands to fiber optic transmission nodes within the non-interested area to obtain low-frequency sampling data, and merging the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset, specifically includes: A configuration instruction containing a first sampling frequency is sent to each optical fiber transmission node within the dynamic monitoring area to receive the performance indicators, alarm logs, and topology connection status reported by the optical fiber transmission node at the first sampling frequency, which serve as the high-frequency sampling data. A configuration instruction containing a second sampling frequency is sent to each fiber optic transmission node in the non-interest area to receive the performance indicators reported by the fiber optic transmission node at the second sampling frequency as the low-frequency sampling data, wherein the second sampling frequency is lower than the first sampling frequency; The high-frequency sampling data and the low-frequency sampling data are aligned by timestamp and then concatenated to generate the multi-level sampling dataset.

[0010] Optionally, the step of predicting the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value for future times, and updating the boundary of the dynamic interest region based on the predicted fiber optic link failure tendency value to generate a predictive dynamic interest region, specifically includes: The time series of the fiber optic link failure tendency values ​​of each fiber optic transmission node is input into the internally configured lightweight gated loop unit to calculate the predicted fiber optic link failure tendency values ​​for several future time steps. A dynamic threshold is calculated based on the statistical characteristics of the fiber optic link failure tendency value of all fiber optic transmission nodes in the network, and fiber optic transmission nodes whose predicted fiber optic link failure tendency value is greater than the dynamic threshold are marked as potentially risky fiber optic transmission nodes. The potentially risky fiber optic transmission nodes are added to the seed node set, and the dynamic interest region delineation step is re-executed based on the updated seed node set to update the region boundary and generate the predictive dynamic interest region.

[0011] Optionally, the step of extracting temporal dynamic features from the high-frequency sampled data acquired within the predictive dynamic attention region to generate a first feature vector specifically includes: The high-frequency sampling data of each optical fiber transmission node within the predictive dynamic interest area are arranged in chronological order to construct a multidimensional time series matrix. The multidimensional time series matrix is ​​input into a long short-term memory network, and through the cyclic processing of its internal forget gate, input gate and output gate, the nonlinear dependency of the state of the optical fiber transmission node over time is extracted. The hidden state vector of the Long Short-Term Memory network at the last time step is output as the first feature vector.

[0012] Optionally, the step of extracting textual semantic features from the low-frequency sampled data obtained from the non-interested region and generating a second feature vector specifically includes: The alarm text data of the non-interested areas are segmented and word-embedded, and then converted into a two-dimensional word vector matrix. The two-dimensional word vector matrix is ​​input into a convolutional neural network, and the local spatial features of key alarm phrases are extracted by the sliding scan of the convolutional kernel. After performing pooling dimensionality reduction on the local spatial features, the second feature vector is generated.

[0013] Optionally, the step of inputting the initial fiber optic transmission node feature vector into a preset time-series graph attention network, and modulating the attention weights between fiber optic transmission nodes in the time-series graph attention network according to the sampling density differences of the regions to which the fiber optic transmission nodes belong, specifically includes: Construct the current network topology graph, and assign an initial fiber transmission node feature vector to each fiber transmission node, which is generated by splicing the first feature vector, the second feature vector, and the numerical sequence representing the changes in the regional membership of the fiber transmission node. In the graph attention layer of the time-series graph attention network, the regional affiliation status of fiber transmission node pairs is identified. For fiber transmission node pairs that are both located in the same dynamic attention region, their attention weights are multiplied by a first gain coefficient. For fiber transmission node pairs that cross regional boundaries, their attention weights are multiplied by a second attenuation coefficient. The high-dimensional fused feature vector is output through the superposition and aggregation of multiple graph attention layers and feature propagation.

[0014] Optionally, the step of inputting the high-dimensional fused feature vector into a classifier for decoding, outputting a diagnostic result including the fault type and the location of the faulty fiber transmission node, and feeding back the location of the faulty fiber transmission node into the calculation step of the fiber link fault tendency value to dynamically update the weighted summation weight parameters specifically includes: The high-dimensional fused feature vector is input into a network branch containing a fully connected layer and a Softmax function. The probability distribution of each fault category is calculated, and the category with the highest probability is selected as the fault type. The high-dimensional fused feature vector is input into the fiber optic transmission node positioning branch network, and the probability vector of each fiber optic transmission node being diagnosed as faulty is calculated and output. The fiber optic transmission nodes with probability values ​​exceeding the positioning threshold are selected as the location of the faulty fiber optic transmission nodes. In the step of calculating the faulty fiber transmission node location, the weighting coefficient of the historical optical attenuation degradation rate corresponding to the diagnosed faulty fiber transmission node is increased. At the same time, the dynamic radius reference value corresponding to the area that repeatedly triggers false alarms is attenuated by introducing a suppression coefficient with a value less than 1.

[0015] Secondly, a feature fusion-based intelligent fault diagnosis system for optical fiber communication links is provided, applied to a fault diagnosis terminal. The system includes: The tendency value calculation module is used to obtain the historical optical attenuation degradation rate, current optical channel load and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node by weighted summing of the historical optical attenuation degradation rate, the current optical channel load and the crosstalk propagation probability of adjacent optical nodes. The dynamic region construction module is used to select high-risk nodes as seed nodes from all optical fiber transmission nodes according to the optical fiber link failure tendency value, and to delineate dynamic interest regions based on the seed nodes and their proximity connections in the network topology, generating region attribute data containing the division status of optical fiber transmission nodes. A multi-level sampling scheduling module is used to send high-frequency sampling instructions to optical fiber transmission nodes in the dynamically concerned area to obtain high-frequency sampling data, and to send low-frequency sampling instructions to optical fiber transmission nodes in the non-concerned area to obtain low-frequency sampling data, and to merge the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset. The prediction and boundary update module is used to extract the fiber optic link failure tendency value generated by each fiber optic transmission node in a continuous historical time step to construct a fiber optic link failure tendency value time series, predict the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value at future time, and update the boundary of the dynamic attention region according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. The multi-dimensional feature extraction module is used to extract the temporal dynamic features from the high-frequency sampled data obtained in the predictive dynamic attention region to generate a first feature vector, and to extract the textual semantic features from the low-frequency sampled data obtained in the non-attention region to generate a second feature vector. The feature fusion and network mapping module is used to concatenate the first feature vector, the second feature vector, and the numerical sequence representing the changes in the identity of members in the region of the optical fiber transmission node to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The closed-loop diagnostic output module is used to input the high-dimensional fused feature vector into the classifier for decoding, output diagnostic results including fault type and faulty fiber transmission node location, and feed back the faulty fiber transmission node location to the tendency value calculation module to dynamically update the weight parameters of the weighted summation.

[0016] Thirdly, an electronic device is provided, comprising: a processor and a memory; the memory is used to store a computer program, which, when executed by the processor, causes the electronic device to perform the feature fusion-based intelligent diagnosis method and system for optical fiber communication link faults described in the first aspect.

[0017] In one possible design, the electronic device described in the third aspect may further include a transceiver. This transceiver may be a transceiver circuit or an interface circuit. The transceiver can be used for communication between the electronic device described in the third aspect and other electronic devices.

[0018] In the embodiments of the present invention, the electronic device described in the third aspect may be a terminal, or a chip (system) or other component or assembly disposed in the terminal, or a system containing the terminal.

[0019] Fourthly, a computer-readable storage medium is provided, comprising: a computer program or instructions; when the computer program or instructions are executed on a computer, the computer causes the computer to perform the feature fusion-based intelligent diagnosis method and system for optical fiber communication link faults described in the first aspect.

[0020] In summary, the above methods and systems have the following technical effects: 1. This invention improves the timeliness of optical fiber physical layer fault diagnosis by combining predictive dynamic sampling with adaptive sampling density fusion. Instead of passively waiting for optical signal loss faults to occur, it uses a time-series prediction model to predict high-risk optical transmission nodes in advance and proactively adjust the dynamic focus area. This achieves early coverage and high-frequency monitoring of potential optical layer fault points, capturing weak precursors to optical fiber physical faults, such as optical signal-to-noise ratio degradation and optical attenuation fluctuations, thus gaining response time for the scheduling of optical network physical routes.

[0021] 2. This invention improves the accuracy of underlying optical channel fault diagnosis by constructing an adaptive fusion mechanism based on sensing sampling density. During feature fusion, the temporal graph attention network dynamically adjusts the attention weights between optical transmission nodes based on whether the optical node data originates from a high-frequency or low-frequency sampling region. This allows the physical features of the optical layer from a high-reliability data source to dominate the fusion decision, effectively suppressing noise interference from sparse data, generating higher-quality high-dimensional fusion feature vectors, and improving the accuracy of optical transmission fault type classification and physical root cause node localization.

[0022] 3. This invention improves the adaptive capability of the entire diagnostic system by establishing a closed-loop feedback channel from diagnostic results to optical layer data sampling strategies. Based on historical optical link diagnostic results, the system can automatically increase the monitoring frequency of faulty optical nodes while reducing resource investment in areas with frequent false alarms. This allows the system to maintain stable performance in the face of dynamically changing fiber optic topologies, achieving a dynamic balance between diagnostic accuracy and optical node maintenance costs. Attached Figure Description

[0023] Figure 1 A flowchart illustrating the intelligent diagnosis method for optical fiber communication link faults based on feature fusion provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the framework of an intelligent diagnostic system for optical fiber communication link faults based on feature fusion, provided in an embodiment of the present invention. Detailed Implementation

[0024] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0025] In this embodiment of the invention, "instruction" can include direct and indirect instructions, as well as explicit and implicit instructions. The information indicated by a certain piece of information is called the information to be instructed. In specific implementation, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly instruct the information to be instructed by instructing other information, where there is a correlation between the other information and the information to be instructed. It can also instruct only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent. Simultaneously, common parts of various pieces of information can be identified and uniformly indicated to reduce the instruction overhead caused by individually indicating the same information.

[0026] Furthermore, the specific indication method can also be any existing indication method, such as, but not limited to, the above-mentioned indication methods and their various combinations. Specific details of various indication methods can be found in existing technologies, and will not be elaborated upon here. As described above, for example, when multiple pieces of information of the same type need to be indicated, the indication methods for different pieces of information may differ. In specific implementation, the required indication method can be selected according to specific needs. This embodiment of the invention does not limit the selected indication method; therefore, the indication methods involved in this embodiment of the invention should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated.

[0027] It should be understood that the information to be indicated can be sent as a whole or divided into multiple sub-information messages sent separately, and the sending period and / or timing of these sub-information messages can be the same or different. The specific sending method is not limited in this embodiment of the invention. The sending period and / or timing of these sub-information messages can be predefined, for example, according to a protocol, or configured by the sending device by sending configuration information to the receiving device.

[0028] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This embodiment of the invention does not limit the specific implementation method. "Saving" can refer to saving in one or more memories. These memories can be separate installations or integrated into the encoder, decoder, processor, or electronic device. Alternatively, some memories can be separately installed, while others are integrated into the decoder, processor, or electronic device. The type of memory can be any form of storage medium, and this embodiment of the invention does not limit this.

[0029] In the embodiments of this invention, "protocol" may refer to a protocol family in the field of communication, a standard protocol with a similar protocol family frame structure, or a related protocol applied to a future intelligent diagnosis method and system for optical fiber communication link faults based on feature fusion. The embodiments of this invention do not specifically limit this.

[0030] In this embodiment of the invention, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.

[0031] In the description of the embodiments of the present invention, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in the embodiments of the present invention is merely a description of the relationship between the related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of the embodiments of the present invention, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of the present invention, the terms "first" and "second" are used in the embodiments of the present invention to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., are not necessarily different. Furthermore, in the embodiments of this invention, words such as "exemplary" or "for example" are used to indicate that something is being described as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this invention should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.

[0032] The network architecture and business scenarios described in the embodiments of this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.

[0033] Figure 1 This is a flowchart illustrating the method provided in an embodiment of the present invention. This intelligent fault diagnosis method for optical fiber communication links based on feature fusion can be applied to fault diagnosis terminals.

[0034] The fault diagnosis terminal can be a terminal device with transmitting, receiving, and processing functions, or a chip or chip system that can be installed in the terminal device. This terminal device can also be referred to as user equipment (UE), access terminal equipment, subscriber unit, user station, mobile station (MS), mobile station, remote station, remote terminal equipment, mobile device, user terminal equipment, terminal equipment, wireless communication equipment, user agent, or user apparatus. The terminal devices in the embodiments of this application may be mobile phones, cellular phones, smartphones, tablets, wireless data cards, personal digital assistants (PDAs), wireless modems, handsets, laptop computers, machine-type communication (MTC) terminal devices, computers with wireless transceiver capabilities, virtual reality (VR) terminal devices, augmented reality (AR) terminal devices, wireless terminal devices in industrial control, wireless terminal devices in self-driving, wireless terminal devices in remote medical care, wireless terminal devices in smart grids, wireless terminal devices in transportation safety, wireless terminal devices in smart cities, wireless terminal devices in smart homes, vehicle-mounted terminal devices, roadside units (RSUs) with terminal device functions, etc. The terminal device in this application may also be an on-board module, on-board component, on-board chip, or on-board unit that is built into a vehicle as one or more components or units. Alternatively, the terminal device may also be a customer-premises equipment (CPE).

[0035] The specific process is as follows: The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node are obtained from the optical fiber communication link. The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes are weighted and summed to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node. Based on the fiber optic link failure tendency value, high-risk nodes are selected from all fiber optic transmission nodes as seed nodes, and dynamic interest areas are delineated based on the seed nodes and their proximity connections in the network topology, generating regional attribute data containing the division status of fiber optic transmission nodes. A high-frequency sampling command is sent to the optical fiber transmission nodes in the dynamic interest area to obtain high-frequency sampling data, and a low-frequency sampling command is sent to the optical fiber transmission nodes in the non-interest area to obtain low-frequency sampling data. The high-frequency sampling data and the low-frequency sampling data are then merged to generate a multi-level sampling dataset. The fiber optic link failure tendency value generated by each fiber optic transmission node within a continuous historical time step is extracted to construct a fiber optic link failure tendency value time series. The fiber optic link failure tendency value time series is predicted to obtain the predicted fiber optic link failure tendency value for future time. The boundary of the dynamic attention region is updated according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. Extract the temporal dynamic features from the high-frequency sampled data obtained within the predictive dynamic attention region to generate a first feature vector, and extract the textual semantic features from the low-frequency sampled data obtained within the non-attention region to generate a second feature vector. The first feature vector, the second feature vector, and the numerical sequence representing the change of the identity of the members in the region of the optical fiber transmission node are concatenated to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The high-dimensional fused feature vector is input into the classifier for decoding, and the output includes the fault type and the location of the faulty fiber transmission node. The location of the faulty fiber transmission node is then fed back into the calculation step of the fiber link fault tendency value to dynamically update the weight parameters of the weighted summation.

[0036] Further, the steps of obtaining the historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and calculating and generating the optical fiber link fault tendency value of each optical fiber transmission node by weighted summation of the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes, specifically include: Receive the number of faults and total running time reported by each optical fiber transmission node within a specified time window, and calculate the ratio between the two to generate the historical optical attenuation degradation rate. Receive the instantaneous bandwidth occupancy rate and queue length reported by each optical fiber transmission node, and perform a linear combination to generate the current optical channel load; Obtain the fault transfer status of each optical fiber transmission node and its neighboring optical fiber transmission nodes in adjacent time steps, and calculate the crosstalk propagation probability of the adjacent optical nodes. The optical fiber link fault tendency value is generated by multiplying the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes by dynamically adjustable first preset weights, second preset weights, and third preset weights, respectively.

[0037] To comprehensively reflect the long-term stability, real-time stress, and neighborhood risks of fiber optic transmission nodes, a dynamic fiber optic link failure tendency value is generated. Specifically, this involves setting a rolling time window, counting the total number of failures and total online operating time of fiber optic transmission nodes within the window, and normalizing the data to obtain the historical optical attenuation degradation rate H.

[0038] The instantaneous bandwidth occupancy rate and queue length are collected, normalized, and linearly combined to generate the current optical channel load L.

[0039] Based on the network topology and the preset fault transfer state matrix, the transfer probabilities of faulty neighbor fiber transmission nodes are extracted and aggregated to generate the crosstalk propagation probability P of adjacent optical nodes.

[0040] The calculation formula for fusing the above parameters using a weighted model is as follows:

[0041] Where F is the final fiber optic link failure tendency value; , , The first, second, and third preset weights are respectively, and satisfy the following conditions: The system can dynamically adjust the values ​​of the above weights based on the macro-level network conditions, such as traffic surges or security threats. Further, the step of selecting high-risk nodes as seed nodes from all fiber optic transmission nodes based on the fiber optic link failure tendency value, and delineating dynamically concerned regions based on the seed nodes and their proximity connections in the network topology, and generating regional attribute data containing the division status of fiber optic transmission nodes, specifically includes: The fiber optic link failure tendency values ​​of all fiber optic transmission nodes are sorted in descending order, and a specified number of the top-ranked fiber optic transmission nodes are selected as the seed nodes. Calculate a dynamic radius for each seed node that is proportional to its fiber optic link failure tendency value, and delineate candidate regions in the network topology based on the dynamic radius; If there are overlapping fiber optic transmission nodes among multiple candidate regions, the candidate regions containing overlapping fiber optic transmission nodes are merged into the same region to generate the dynamic region of interest. Record the list of internal fiber optic transmission nodes, the list of edge fiber optic transmission nodes, and the timestamp of the region formation in the dynamically monitored area, and combine them to generate the region attribute data.

[0042] Further, the step of issuing high-frequency sampling commands to fiber optic transmission nodes within the dynamically interested area to obtain high-frequency sampling data, and issuing low-frequency sampling commands to fiber optic transmission nodes within the non-interested area to obtain low-frequency sampling data, and merging the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset, specifically includes: A configuration instruction containing a first sampling frequency is sent to each optical fiber transmission node within the dynamic monitoring area to receive the performance indicators, alarm logs, and topology connection status reported by the optical fiber transmission node at the first sampling frequency, which serve as the high-frequency sampling data. A configuration instruction containing a second sampling frequency is sent to each fiber optic transmission node in the non-interest area to receive the performance indicators reported by the fiber optic transmission node at the second sampling frequency as the low-frequency sampling data, wherein the second sampling frequency is lower than the first sampling frequency; The high-frequency sampling data and the low-frequency sampling data are aligned by timestamp and then concatenated to generate the multi-level sampling dataset.

[0043] In step S2, the abstract risk indicators are transformed into regions of interest with physical topological significance. Specifically, this includes sorting the fiber optic link failure tendency values ​​in descending order and selecting the top 10% to 30% of fiber optic transmission nodes as seed nodes.

[0044] Using the seed node as the center, calculate the number of hops that are positively correlated with its fiber optic link failure tendency value as the dynamic radius, and delineate the candidate region.

[0045] Perform a union operation on candidate regions with overlapping fiber optic transmission nodes to output a list of fiber optic transmission nodes for non-overlapping dynamically interested regions and information on boundary fiber optic transmission nodes.

[0046] For fiber optic transmission nodes within the dynamically monitored area, a first sampling frequency with a higher frequency is used to collect performance, logs, and topology status data; for fiber optic transmission nodes in non-monetary areas, a second sampling frequency with a lower frequency is used to collect only basic performance indicators; the two sets of data are aligned by timestamps to generate a multi-level sampling dataset. Further, the steps of predicting the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value at future times, and updating the boundary of the dynamic interest region based on the predicted fiber optic link failure tendency value to generate a predictive dynamic interest region specifically include: The time series of the fiber optic link failure tendency values ​​of each fiber optic transmission node is input into the internally configured lightweight gated loop unit to calculate the predicted fiber optic link failure tendency values ​​for several future time steps. A dynamic threshold is calculated based on the statistical characteristics of the fiber optic link failure tendency value of all fiber optic transmission nodes in the network, and fiber optic transmission nodes whose predicted fiber optic link failure tendency value is greater than the dynamic threshold are marked as potentially risky fiber optic transmission nodes. The potentially risky fiber optic transmission nodes are added to the seed node set, and the dynamic interest region delineation steps are re-executed based on the updated seed node set to update the region boundary and generate the predictive dynamic interest region.

[0047] In step S3, to achieve proactive monitoring, the following is specifically included: Extract the historical fiber optic link failure tendency value sequence for 5 consecutive time steps, input it into a lightweight gated cyclic unit (GRU), and predict the fiber optic link failure tendency value for future time steps.

[0048] The dynamic threshold is calculated based on the real-time mean and standard deviation of the failure tendency value of the entire network's optical fiber links. If the predicted failure tendency value of an optical fiber transmission node is greater than the dynamic threshold, it is marked as a potentially risky optical fiber transmission node and added to the seed node set. Based on this, the boundary is recalculated and expanded to generate a predictive dynamic area of ​​interest.

[0049] Further, the step of extracting temporal dynamic features from the high-frequency sampled data acquired within the predictive dynamic attention region and generating a first feature vector specifically includes: The high-frequency sampling data of each optical fiber transmission node within the predictive dynamic interest area are arranged in chronological order to construct a multidimensional time series matrix. The multidimensional time series matrix is ​​input into a long short-term memory network, and through the cyclic processing of its internal forget gate, input gate and output gate, the nonlinear dependency of the state of the optical fiber transmission node over time is extracted. The hidden state vector of the Long Short-Term Memory network at the last time step is output as the first feature vector.

[0050] Further, the step of extracting textual semantic features from the low-frequency sampled data obtained from the non-interested region and generating a second feature vector specifically includes: The alarm text data of the non-interested areas are segmented and word-embedded, and then converted into a two-dimensional word vector matrix. The two-dimensional word vector matrix is ​​input into a convolutional neural network, and the local spatial features of key alarm phrases are extracted by the sliding scan of the convolutional kernel. After performing pooling dimensionality reduction on the local spatial features, the second feature vector is generated.

[0051] In step S4, temporal dynamics and textual semantic features are extracted respectively, specifically including: A high-frequency multidimensional time series is constructed and input into an LSTM network. By controlling the forget gate, input gate and output gate inside the network, the hidden state vector of the last time step is selected as the first feature vector output.

[0052] The unstructured alarm text collected at low frequency is converted into a two-dimensional matrix through word segmentation and word embedding. The matrix is ​​then input into a CNN network to extract local phrase features. After dimensionality reduction by max pooling, a second feature vector is generated.

[0053] Further, the step of inputting the initial fiber optic transmission node feature vector into a preset time-series graph attention network, and modulating the attention weights between fiber optic transmission nodes in the time-series graph attention network according to the sampling density differences of the regions to which the fiber optic transmission nodes belong, specifically includes: Construct the current network topology graph, and assign an initial fiber transmission node feature vector to each fiber transmission node, which is generated by splicing the first feature vector, the second feature vector, and the numerical sequence representing the changes in the regional membership of the fiber transmission node. In the graph attention layer of the time-series graph attention network, the regional affiliation status of fiber transmission node pairs is identified. For fiber transmission node pairs that are both located in the same dynamic attention region, their attention weights are multiplied by a first gain coefficient. For fiber transmission node pairs that cross regional boundaries, their attention weights are multiplied by a second attenuation coefficient. The high-dimensional fused feature vector is output through the superposition and aggregation of multiple graph attention layers and feature propagation.

[0054] In step S5, topology-aware feature aggregation is performed, specifically including: The first feature vector, the second feature vector, and the numerical sequence of changes in the region boundary are concatenated to form the initial feature vector of the optical fiber transmission node.

[0055] The input is fed into a time-series graph attention network, dynamically modulating the attention weights between connected fiber optic transmission nodes: if fiber optic transmission nodes are within the same dynamic region of interest (high-frequency sampling has high reliability), their attention weights are multiplied by a first gain coefficient greater than 1; if fiber optic transmission nodes cross region boundaries (low-frequency sampling has uncertainty), their attention weights are multiplied by a second attenuation coefficient less than 1. After multi-layer stacking and aggregation, a high-dimensional fused feature vector is output.

[0056] Further, the high-dimensional fused feature vector is input into a classifier for decoding, outputting a diagnostic result containing the fault type and the location of the faulty fiber optic transmission node. The location of the faulty fiber optic transmission node is then fed back into the calculation step of the fiber optic link fault tendency value to dynamically update the weighted summation parameters. This step specifically includes: The high-dimensional fused feature vector is input into a network branch containing a fully connected layer and a Softmax function. The probability distribution of each fault category is calculated, and the category with the highest probability is selected as the fault type. The high-dimensional fused feature vector is input into the fiber optic transmission node positioning branch network, and the probability vector of each fiber optic transmission node being diagnosed as faulty is calculated and output. The fiber optic transmission nodes with probability values ​​exceeding the positioning threshold are selected as the location of the faulty fiber optic transmission nodes. In the step of calculating the faulty fiber transmission node location, the weighting coefficient of the historical optical attenuation degradation rate corresponding to the diagnosed faulty fiber transmission node is increased. At the same time, the dynamic radius reference value corresponding to the area that repeatedly triggers false alarms is attenuated by introducing a suppression coefficient with a value less than 1.

[0057] In step S6, the diagnostic conclusion is output and the subsequent prediction model is optimized. Specifically, this includes: inputting the high-dimensional fusion feature vector into the fully connected layer and the Softmax function, and outputting the fault type with the highest probability.

[0058] Simultaneously calculate the root cause confidence of the fiber optic transmission node through the fiber optic transmission node location branch network, and output the location of the faulty fiber optic transmission node with a probability exceeding a preset threshold.

[0059] Feedback mechanism: The system will report back the confirmed faulty fiber optic transmission nodes and increase the weight of their corresponding historical optical attenuation degradation rate $\alpha$; if a certain area is repeatedly identified as high-risk but the fault cannot be located, a suppression coefficient of less than 1 will be introduced when calculating the dynamic radius of the seed node in that area to achieve adaptive resource allocation.

[0060] The method provided by the embodiments of the present invention has been described in detail above. The following, in conjunction with... Figure 2 A detailed description of an intelligent optical fiber communication link fault diagnosis system based on feature fusion for implementing the methods provided in the embodiments of the present invention, applied to a fault diagnosis terminal, the system comprising: The tendency value calculation module is used to obtain the historical optical attenuation degradation rate, current optical channel load and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node by weighted summing of the historical optical attenuation degradation rate, the current optical channel load and the crosstalk propagation probability of adjacent optical nodes. The dynamic region construction module is used to select high-risk nodes as seed nodes from all optical fiber transmission nodes according to the optical fiber link failure tendency value, and to delineate dynamic interest regions based on the seed nodes and their proximity connections in the network topology, generating region attribute data containing the division status of optical fiber transmission nodes. A multi-level sampling scheduling module is used to send high-frequency sampling instructions to optical fiber transmission nodes in the dynamically concerned area to obtain high-frequency sampling data, and to send low-frequency sampling instructions to optical fiber transmission nodes in the non-concerned area to obtain low-frequency sampling data, and to merge the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset. The prediction and boundary update module is used to extract the fiber optic link failure tendency value generated by each fiber optic transmission node in a continuous historical time step to construct a fiber optic link failure tendency value time series, predict the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value at future time, and update the boundary of the dynamic attention region according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. The multi-dimensional feature extraction module is used to extract the temporal dynamic features from the high-frequency sampled data obtained in the predictive dynamic attention region to generate a first feature vector, and to extract the textual semantic features from the low-frequency sampled data obtained in the non-attention region to generate a second feature vector. The feature fusion and network mapping module is used to concatenate the first feature vector, the second feature vector, and the numerical sequence representing the changes in the identity of members in the region of the optical fiber transmission node to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The closed-loop diagnostic output module is used to input the high-dimensional fused feature vector into the classifier for decoding, output diagnostic results including fault type and faulty fiber transmission node location, and feed back the faulty fiber transmission node location to the tendency value calculation module to dynamically update the weight parameters of the weighted summation.

[0061] The electronic device provided in this embodiment of the invention, exemplarily, can be a network device, or a chip (system) or other component or assembly that can be disposed in a network device. The electronic device may include a processor. Optionally, the electronic device may also include a memory and / or a transceiver. The processor is coupled to the memory and transceiver, for example, by means of a communication bus connection.

[0062] The following is a detailed introduction to the various components of the electronic device: In this context, the processor is the control center of the electronic device. It can be a single processor or a collective term for multiple processing elements. For example, a processor can be one or more central processing units (CPUs), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention, such as one or more digital signal processors (DSPs), or one or more field-programmable gate arrays (FPGAs).

[0063] Alternatively, the processor can perform various functions of the electronic device, such as the methods described above, by running or executing software programs stored in memory and by calling data stored in memory.

[0064] In a specific implementation, as one example, the processor may include one or more CPUs, such as CPU0 and CPU1.

[0065] In a specific implementation, as one example, the electronic device may also include multiple processors. Each of these processors may be a single-core processor (single-CPU) or a multi-core processor (multi-CPU). Here, a processor may refer to one or more devices, circuits, and / or processing cores used to process data (e.g., computer program instructions).

[0066] The memory is used to store the software program that executes the solution of the present invention, and the execution is controlled by the processor. The specific implementation method can be referred to the above method embodiment, and will not be repeated here.

[0067] Optionally, the memory can be read-only memory (ROM) or other types of static storage devices capable of storing static information and instructions, random access memory (RAM) or other types of dynamic storage devices capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto. The memory can be integrated with the processor or exist independently and coupled to the processor through the interface circuit of the electronic device; the embodiments of the present invention do not specifically limit this.

[0068] A transceiver is used for communication with other electronic devices. For example, if the electronic device is a terminal, the transceiver can be used to communicate with a network device or with another terminal device. Similarly, if the electronic device is a network device, the transceiver can be used to communicate with a terminal or with another network device.

[0069] Optionally, the transceiver may include a receiver and a transmitter. The receiver is used to implement the receiving function, and the transmitter is used to implement the sending function.

[0070] Optionally, the transceiver can be integrated with the processor or exist independently and coupled to the processor through the interface circuit of the electronic device. This embodiment of the invention does not specifically limit this.

[0071] It is understood that the structure of the electronic device in this embodiment does not constitute a limitation on the electronic device. The actual electronic device may include more or fewer components, or combine certain components, or have different component arrangements.

[0072] Furthermore, the technical effects of the electronic device can be referred to the technical effects of the method described in the above method embodiments, and will not be repeated here.

[0073] It should be understood that the processor in the embodiments of the present invention can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

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

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

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

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

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

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

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

[0081] In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0082] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0083] In addition, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0084] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0085] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for intelligent diagnosis of optical fiber communication link faults based on feature fusion, characterized in that, Applied to a fault diagnosis terminal, the method includes the following steps: The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node are obtained from the optical fiber communication link. The historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes are weighted and summed to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node. Based on the fiber optic link failure tendency value, high-risk nodes are selected from all fiber optic transmission nodes as seed nodes, and dynamic interest areas are delineated based on the seed nodes and their proximity connections in the network topology, generating regional attribute data containing the division status of fiber optic transmission nodes. A high-frequency sampling command is sent to the optical fiber transmission nodes in the dynamic interest area to obtain high-frequency sampling data, and a low-frequency sampling command is sent to the optical fiber transmission nodes in the non-interest area to obtain low-frequency sampling data. The high-frequency sampling data and the low-frequency sampling data are then merged to generate a multi-level sampling dataset. The fiber optic link failure tendency value generated by each fiber optic transmission node within a continuous historical time step is extracted to construct a fiber optic link failure tendency value time series. The fiber optic link failure tendency value time series is predicted to obtain the predicted fiber optic link failure tendency value for future time. The boundary of the dynamic attention region is updated according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. Extract the temporal dynamic features from the high-frequency sampled data obtained within the predictive dynamic attention region to generate a first feature vector, and extract the textual semantic features from the low-frequency sampled data obtained within the non-attention region to generate a second feature vector. The first feature vector, the second feature vector, and the numerical sequence representing the change of the identity of the members in the region of the optical fiber transmission node are concatenated to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The high-dimensional fused feature vector is input into the classifier for decoding, and the output includes the fault type and the location of the faulty fiber transmission node. The location of the faulty fiber transmission node is then fed back into the calculation step of the fiber link fault tendency value to dynamically update the weight parameters of the weighted summation.

2. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The steps of obtaining the historical optical attenuation degradation rate, current optical channel load, and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and calculating the optical link fault tendency value of each optical fiber transmission node by weighted summation of the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes, specifically include: Receive the number of faults and total running time reported by each optical fiber transmission node within a specified time window, and calculate the ratio between the two to generate the historical optical attenuation degradation rate. Receive the instantaneous bandwidth occupancy rate and queue length reported by each optical fiber transmission node, and perform a linear combination to generate the current optical channel load; Obtain the fault transfer status of each optical fiber transmission node and its neighboring optical fiber transmission nodes in adjacent time steps, and calculate the crosstalk propagation probability of the adjacent optical nodes. The optical fiber link fault tendency value is generated by multiplying the historical optical attenuation degradation rate, the current optical channel load, and the crosstalk propagation probability of adjacent optical nodes by dynamically adjustable first preset weights, second preset weights, and third preset weights, respectively.

3. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The steps of selecting high-risk nodes as seed nodes from all fiber optic transmission nodes based on the fiber optic link failure tendency value, and delineating dynamically concerned regions based on the seed nodes and their proximity connections in the network topology, and generating region attribute data containing the division status of fiber optic transmission nodes, specifically include: The fiber optic link failure tendency values ​​of all fiber optic transmission nodes are sorted in descending order, and the top 10% to 30% of fiber optic transmission nodes are selected as seed nodes. A dynamic radius proportional to the fiber optic link failure tendency value is calculated for each seed node, and candidate regions are delineated in the network topology based on the dynamic radius. If there are overlapping fiber optic transmission nodes among multiple candidate regions, the candidate regions containing overlapping fiber optic transmission nodes are merged into the same region to generate the dynamic region of interest. Record the list of internal fiber optic transmission nodes, the list of edge fiber optic transmission nodes, and the timestamp of the region formation in the dynamically monitored area, and combine them to generate the region attribute data.

4. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The steps of issuing high-frequency sampling commands to fiber optic transmission nodes within the dynamically monitored area to obtain high-frequency sampling data, and issuing low-frequency sampling commands to fiber optic transmission nodes within the non-monitored area to obtain low-frequency sampling data, and merging the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset, specifically include: A configuration instruction containing a first sampling frequency is sent to each optical fiber transmission node within the dynamic monitoring area to receive the performance indicators, alarm logs, and topology connection status reported by the optical fiber transmission node at the first sampling frequency, which serve as the high-frequency sampling data. A configuration instruction containing a second sampling frequency is sent to each fiber optic transmission node in the non-interest area to receive the performance indicators reported by the fiber optic transmission node at the second sampling frequency as the low-frequency sampling data, wherein the second sampling frequency is lower than the first sampling frequency; The high-frequency sampling data and the low-frequency sampling data are aligned by timestamp and then concatenated to generate the multi-level sampling dataset.

5. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The steps of predicting the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value for future time periods, and updating the boundary of the dynamically concerned region based on the predicted fiber optic link failure tendency value to generate a predictive dynamically concerned region, specifically include: The time series of the fiber optic link failure tendency values ​​of each fiber optic transmission node is input into the internally configured lightweight gated loop unit to calculate the predicted fiber optic link failure tendency value for future time steps. A dynamic threshold is calculated based on the statistical characteristics of the fiber optic link failure tendency value of all fiber optic transmission nodes in the network, and fiber optic transmission nodes whose predicted fiber optic link failure tendency value is greater than the dynamic threshold are marked as potentially risky fiber optic transmission nodes. The potentially risky fiber optic transmission nodes are added to the seed node set, and the dynamic interest region delineation step is re-executed based on the updated seed node set to update the region boundary and generate the predictive dynamic interest region.

6. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The step of extracting temporal dynamic features from the high-frequency sampled data acquired within the predictive dynamic interest region and generating a first feature vector specifically includes: The high-frequency sampling data of each optical fiber transmission node within the predictive dynamic interest area are arranged in chronological order to construct a multidimensional time series matrix. The multidimensional time series matrix is ​​input into a long short-term memory network, and through the cyclic processing of its internal forget gate, input gate and output gate, the nonlinear dependency of the state of the optical fiber transmission node over time is extracted. The hidden state vector of the Long Short-Term Memory network at the last time step is output as the first feature vector.

7. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The step of extracting textual semantic features from the low-frequency sampled data obtained from the non-interested region and generating a second feature vector specifically includes: The alarm text data of the non-interested areas are segmented and word-embedded, and then converted into a two-dimensional word vector matrix. The two-dimensional word vector matrix is ​​input into a convolutional neural network, and the local spatial features of key alarm phrases are extracted by the sliding scan of the convolutional kernel. After performing pooling dimensionality reduction on the local spatial features, the second feature vector is generated.

8. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The steps of inputting the initial fiber optic transmission node feature vector into a preset time-series graph attention network, and modulating the attention weights between fiber optic transmission nodes in the time-series graph attention network according to the sampling density differences of the regions to which the fiber optic transmission nodes belong, specifically include: Construct the current network topology graph, and assign an initial fiber transmission node feature vector to each fiber transmission node, which is generated by splicing the first feature vector, the second feature vector, and the numerical sequence representing the changes in the regional membership of the fiber transmission node. In the graph attention layer of the time-series graph attention network, the region affiliation status of fiber transmission node pairs is identified. For fiber transmission node pairs that are both located in the same dynamic attention region, their attention weights are multiplied by a first gain coefficient greater than 1. For fiber transmission node pairs that cross region boundaries, their attention weights are multiplied by a second attenuation coefficient less than 1. The high-dimensional fused feature vector is output through the superposition and aggregation of multiple graph attention layers and feature propagation.

9. The intelligent diagnosis method for optical fiber communication link faults based on feature fusion according to claim 1, characterized in that, The high-dimensional fused feature vector is input into a classifier for decoding, and a diagnostic result containing the fault type and location of the faulty fiber optic transmission node is output. The location of the faulty fiber optic transmission node is then fed back into the calculation step of the fiber optic link fault tendency value to dynamically update the weighted summation weight parameters. This step specifically includes: The high-dimensional fused feature vector is input into a network branch containing a fully connected layer and a Softmax function. The probability distribution of each fault category is calculated, and the category with the highest probability is selected as the fault type. The high-dimensional fused feature vector is input into the fiber optic transmission node positioning branch network, and the probability vector of each fiber optic transmission node being diagnosed as faulty is calculated and output. The fiber optic transmission nodes with probability values ​​exceeding the positioning threshold are selected as the location of the faulty fiber optic transmission nodes. In the step of calculating the faulty fiber transmission node location, the weighting coefficient of the historical optical attenuation degradation rate corresponding to the diagnosed faulty fiber transmission node is increased. At the same time, the dynamic radius reference value corresponding to the area that repeatedly triggers false alarms is attenuated by introducing a suppression coefficient with a value less than 1.

10. A smart fault diagnosis system for optical fiber communication links based on feature fusion, characterized in that, The system, applied to a fault diagnosis terminal, includes: The tendency value calculation module is used to obtain the historical optical attenuation degradation rate, current optical channel load and crosstalk propagation probability of adjacent optical nodes reported by each optical fiber transmission node in the optical fiber communication link, and to calculate and generate the optical fiber link fault tendency value of each optical fiber transmission node by weighted summing of the historical optical attenuation degradation rate, the current optical channel load and the crosstalk propagation probability of adjacent optical nodes. The dynamic region construction module is used to select high-risk nodes as seed nodes from all optical fiber transmission nodes according to the optical fiber link failure tendency value, and to delineate dynamic interest regions based on the seed nodes and their proximity connections in the network topology, generating region attribute data containing the division status of optical fiber transmission nodes. A multi-level sampling scheduling module is used to send high-frequency sampling instructions to optical fiber transmission nodes in the dynamically concerned area to obtain high-frequency sampling data, and to send low-frequency sampling instructions to optical fiber transmission nodes in the non-concerned area to obtain low-frequency sampling data, and to merge the high-frequency sampling data and the low-frequency sampling data to generate a multi-level sampling dataset. The prediction and boundary update module is used to extract the fiber optic link failure tendency value generated by each fiber optic transmission node in a continuous historical time step to construct a fiber optic link failure tendency value time series, predict the fiber optic link failure tendency value time series to obtain the predicted fiber optic link failure tendency value at future time, and update the boundary of the dynamic attention region according to the predicted fiber optic link failure tendency value to generate a predictive dynamic attention region. The multi-dimensional feature extraction module is used to extract the temporal dynamic features from the high-frequency sampled data obtained in the predictive dynamic attention region to generate a first feature vector, and to extract the textual semantic features from the low-frequency sampled data obtained in the non-attention region to generate a second feature vector. The feature fusion and network mapping module is used to concatenate the first feature vector, the second feature vector, and the numerical sequence representing the changes in the identity of members in the region of the optical fiber transmission node to generate an initial optical fiber transmission node feature vector. The initial optical fiber transmission node feature vector is input into a preset time-series graph attention network, and the attention weights between optical fiber transmission nodes in the time-series graph attention network are modulated according to the sampling density difference of the region to which the optical fiber transmission node belongs, and a high-dimensional fused feature vector is output. The closed-loop diagnostic output module is used to input the high-dimensional fused feature vector into the classifier for decoding, output diagnostic results including fault type and faulty fiber transmission node location, and feed back the faulty fiber transmission node location to the tendency value calculation module to dynamically update the weight parameters of the weighted summation.

Images