Automatic testing method for FTTR optical link

Abstract

The invention relates to an automatic test method of an FTTR optical link, which comprises the following steps: acquiring specified parameters of the FTTR optical link to be detected, establishing communication connection between a test control center and FTTR equipment, dynamically generating an optical link test instruction and sending the optical link test instruction to the FTTR equipment, and testing the optical link according to the received optical link test instruction. The test control center performs comparative analysis on the received test data and preset standard data, inputs the test data into a machine learning model for training and updating, judges whether the optical link has a fault based on a comparative analysis result and a prediction result, and marks a fault type and a fault position by using a fault positioning algorithm if the optical link has the fault; and generating a test report based on the comparative analysis result, the fault judgment result, the fault type and the fault position, and automatically generating a fault solution according to the fault type. According to the method, automatic and intelligent testing of the FTTR optical link is realized, and the fault diagnosis efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of optical fiber communication, and in particular relates to an automatic testing method for an FTTR optical link. Background Art

[0002] In recent years, with the rapid development of communications technology, fiber-to-the-room (FTTR) technology has been widely adopted in home networks, enterprise office networks, and public area network coverage due to its high bandwidth, low latency, and stable transmission performance. As the core infrastructure for data transmission, the stability and reliability of FTTR optical links directly impact network service quality and user experience. However, in practical applications, FTTR optical links face a variety of potential problems, such as fiber aging, optical module failures, and abnormal link connections. These failures can lead to adverse consequences such as network speed degradation and signal interruption. Therefore, efficient and accurate testing and troubleshooting of FTTR optical links are particularly important.

[0003] Traditional FTTR optical link testing methods rely primarily on manual operation, requiring technicians to carry specialized testing equipment and individually check each node and parameter of the optical link. This approach is not only inefficient but also susceptible to human error, resulting in inaccurate test results and imprecise fault location. Furthermore, due to the increasingly complex network topologies of FTTR optical links, as well as the diverse link lengths and optical module models, traditional methods struggle to adapt to the testing requirements of diverse scenarios. With the continuous increase in the number of FTTR devices and the continued expansion of network scale, the cost and difficulty of manual testing have also increased exponentially, making it unable to meet the demands of the rapid development of modern communication networks.

[0004] With the continuous advancement of artificial intelligence and machine learning technologies, their application to optical link testing has become an important approach to addressing the aforementioned issues. By using machine learning models to learn and analyze large amounts of historical data, it is possible to uncover the potential relationships between optical link faults and various parameters, enabling intelligent fault prediction and diagnosis. However, existing machine learning-based optical link testing solutions still have shortcomings in data processing, test instruction generation, and fault location. These solutions are unable to fully utilize the specified parameters of FTTR optical links for dynamic and precise testing, making it difficult to quickly and accurately generate effective fault resolution solutions.

[0005] Therefore, there is an urgent need to develop a method that can integrate multiple parameter information, realize automated and intelligent testing of FTTR optical links, and quickly locate faults and provide solutions to ensure the stable operation of FTTR optical links and efficient services of communication networks. Summary of the Invention

[0006] The purpose of the present invention is to provide an automated testing method for FTTR optical links, which is used to realize automated and intelligent testing of FTTR optical links, and to quickly locate faults and provide solutions, so as to ensure the stable operation of FTTR optical links and the efficient service of communication networks.

[0007] In order to solve the above technical problems, the technical solutions adopted by the present invention are as follows:

[0008] An automated testing method for an FTTR optical link comprises the following steps:

[0009] S1: Collect the specified parameters of the FTTR optical link to be tested, including link length, optical module model, and network topology, and establish a communication connection between the test control center and the FTTR equipment;

[0010] S2: The test control center dynamically generates an optical link test instruction according to the specified parameters based on a built-in machine learning model, and sends the optical link test instruction to the FTTR device;

[0011] S3: The FTTR device controls the corresponding test module to test the optical link according to the received optical link test instruction, and feeds back the test data to the test control center;

[0012] S4: The test control center compares and analyzes the received test data with preset standard data, and inputs the test data into the machine learning model for training and updating;

[0013] S5: The test control center determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If there is a fault, the fault type and fault location are marked using a fault location algorithm.

[0014] S6: Generate a test report based on the comparison and analysis results, fault judgment results, fault type and fault location, and automatically generate a fault solution according to the fault type.

[0015] Preferably, the specific process of establishing the communication connection between the test control center and the FTTR device in step S1 is as follows:

[0016] S11: Start the test control center and send a device scanning signal to the network through the preset communication port. Use broadcast or multicast to detect the FTTR device to be tested. After receiving the scanning signal, the FTTR device to be tested returns a response data packet containing the device's unique identifier and a list of supported communication protocols.

[0017] S12: The test control center receives and parses the response data packet, and selects a communication protocol with the highest priority and supported by both parties from its supported protocol library according to the communication protocol list fed back by the FTTR device to be tested, and sends a protocol negotiation request to the FTTR device;

[0018] S13: The test control center performs identity authentication with the FTTR device. The test control center sends an encrypted authentication request to the FTTR device. The FTTR device decrypts the request using the pre-shared key and verifies the legitimacy of the request. If the authentication succeeds, an authentication success response is returned, and step S14 is executed.

[0019] S14: Configure connection parameters. The test control center negotiates with the FTTR device to determine the connection timeout, retransmission mechanism, and data cache size, allocates a port number for communication, and establishes a data transmission channel.

[0020] Preferably, in step S2, the specific process of the test control center dynamically generating the optical link test instruction according to the specified parameters based on the built-in machine learning model is as follows:

[0021] S21: Integrate the specified parameters to form a structured data set, and convert the data in the structured data set by label encoding;

[0022] S22: performing feature extraction on the converted structured data set, screening out key features that are strongly correlated with the optical link failure to obtain strongly correlated features;

[0023] S23: Inputting the strong correlation features into a built-in machine learning model, using the model to identify patterns and regularities in the data, and predicting the type of optical link faults and the peak fault-incidence period under specified link lengths, optical module models, and environmental conditions;

[0024] S24: The test control center formulates an optical link test instruction strategy based on the prediction results, generates specific optical link test instructions based on the formulated test instruction strategy, and optimizes the instructions in combination with the performance and resource limitations of the FTTR equipment to be tested.

[0025] S25: Verify the optical link test instruction by simulating the execution process of the instruction on the FTTR device to be tested to check the integrity, accuracy and feasibility of the instruction.

[0026] Preferably, in step S4, the test control center compares and analyzes the received test data with the preset standard data, and inputs the test data into the machine learning model for training and updating. The specific process is as follows:

[0027] S41: The test control center removes duplicate data from the received test data, determines data uniqueness through timestamps and device identifiers, checks data integrity, completes missing data, and converts the data into a format recognizable by the machine learning model.

[0028] S42: Compare and analyze the processed test data with preset standard data, generate a preliminary fault diagnosis result based on the comparison and analysis results, mark the optical link with possible faults and preliminary fault type diagnosis information;

[0029] S43: Feature extraction is performed on the test data. For optical power data, features such as power fluctuation amplitude and power change trend are extracted. For bit error data, features such as bit error rate change rate and bit error distribution pattern are extracted. A specified feature selection algorithm is used to screen out key features that are highly correlated with optical link faults, remove redundant and irrelevant features, and reduce data dimensionality. The preliminary fault judgment results are used as label data and integrated with the extracted feature data to form a data set for model training.

[0030] S44: Divide the integrated data set into a training set and a validation set, randomly divide them according to the specified ratio, use the training set to train the built-in machine learning model, and during the training process, use the cross-validation method to adjust the hyperparameters. Use the validation set to evaluate the model and calculate the model's specified evaluation indicators. If the evaluation result does not meet the preset standard, readjust the hyperparameters and train and evaluate again until the model performance meets the requirements. Save the updated model for subsequent test instruction generation and fault prediction.

[0031] Preferably, in step S5, the test control center determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If there is a fault, the specific process of marking the fault type and fault location using the fault location algorithm is as follows:

[0032] S51: The test control center integrates the comparative analysis results with the prediction results. The comparative analysis results include the deviation of optical power, bit error rate, link connectivity indicators from standard data and preliminary fault judgment. The model prediction results are based on the prediction of the current optical link status after historical data training. The two types of results are integrated according to the specified key information to form a unified judgment data set.

[0033] S52: Perform preliminary fault diagnosis based on the fused data set. For optical power, if the comparative analysis results show that the power exceeds the standard range and the machine learning model also predicts a power anomaly risk, a preliminary judgment is made that a fault exists in the optical power-related link. For bit error rate, if the actual bit error rate is higher than the threshold and the model predicts an upward trend in the bit error rate, a bit error problem is marked. For link connectivity, if the comparative analysis confirms a link interruption and the model predicts an unstable connection, a link connectivity fault is determined. After preliminary judgment, optical links suspected of having faults are screened out.

[0034] S53: For optical links that are initially determined to have faults, the fault type is refined. Using the information about the fault mode in the prediction results and the specific deviations of each indicator during comparative analysis, if the optical power continues to decrease and the model predicts power attenuation due to optical module aging, the fault type is determined to be optical module aging, based on the optical module model and usage time. If the bit error rate fluctuates dramatically and the model indicates excessive optical fiber loss, the link length and environmental factors are considered to determine that the fault type is abnormal optical fiber loss.

[0035] S54: After determining the fault type, a fault location algorithm is used to mark the fault location. For optical power-related faults, if the fault is determined to be an optical module problem, the optical link topology is queried to locate the FTTR device node where the optical module is installed. If the fault is a fiber loss fault, the optical time domain reflectometry algorithm is used to calculate the location of the fiber fault point based on the optical signal reflection time and intensity. For link connectivity faults, the fault is checked step by step from the trunk link to the branch link. Based on the network topology and device connection relationships, the specific link segment or connection node where the fault occurred is determined through segmented testing.

[0036] S55: After completing the marking of the fault type and location, the results are verified, and the test data and historical records are retrieved again to verify the rationality of the fault judgment and location. If there are contradictions or uncertainties, part of the judgment process is re-executed to ensure that the fault type and location information are accurate and form a complete fault marking result.

[0037] Preferably, in step S54, the specific process of calculating the location of the optical fiber fault point according to the reflection time and intensity of the optical signal using the optical time domain reflectometry algorithm is as follows:

[0038] S541: The test control center controls the optical transmitter module in the FTTR device to send a series of optical pulse signals to the optical fiber link under test. The optical pulse signals have a preset wavelength, pulse width, and power.

[0039] S542: When the optical pulse is transmitted in the optical fiber, the optical receiving module in the FTTR device collects the reflected optical signal in real time and converts the optical signal into an electrical signal;

[0040] S543: Preprocessing the collected electrical signal: using an amplifier to amplify the weak electrical signal to enhance the signal strength, using a filter to remove noise and interference in the signal, and converting the analog electrical signal into a digital signal;

[0041] S544: Analyze the pre-processed digital signal and plot a curve of optical power versus time, i.e., an OTDR curve. By analyzing the characteristics of the OTDR curve, the position, height, and width of the reflection peak are identified. The position of the reflection peak corresponds to the round-trip transmission time of the optical signal, the height reflects the intensity of the reflected optical signal, and the width reflects the characteristics of the reflection point.

[0042] S545: Based on the propagation speed of the optical signal in the optical fiber and the round-trip time of the reflected optical signal, the distance from the fault point to the test end is calculated using the formula L = v × t / 2, where L is the distance to the fault point, v is the propagation speed of light in the optical fiber, and t is the time from the optical signal being transmitted to being reflected back to the test end. By measuring the time corresponding to the reflection peak position in the OTDR curve and combining it with the known optical fiber propagation speed, the location of the optical fiber fault point is calculated. The type of fault is determined based on the intensity and shape of the reflection peak.

[0043] S546: Verify the calculated fault point location and correct the calculation result based on the actual fiber optic laying path and the impact of the bending degree on the optical signal transmission time. If there are multiple reflection peaks or complex curve characteristics, eliminate misjudgments by comparing historical OTDR data with the topology information of the fiber optic link, and feedback the fault point location information to the test control center.

[0044] Preferably, in step S6, a test report is generated based on the comparative analysis results, fault judgment results, fault type and fault location, and a specific process of automatically generating a fault solution according to the fault type is as follows:

[0045] S61: The test control center will compare the analysis results, fault judgment results, fault type and fault location data;

[0046] S62: Build the framework structure of the test report according to industry standards and actual application requirements;

[0047] S63: Convert classified and organized data into intuitive visual charts;

[0048] S64: Fill the organized data and visual charts into the corresponding sections according to the established report structure.

[0049] The beneficial effects of the present invention include:

[0050] The automated testing method for FTTR optical links provided by the present invention collects specified parameters of the FTTR optical link to be tested, establishes a communication connection between a test control center and the FTTR equipment, dynamically generates optical link test instructions and sends them to the FTTR equipment. The optical link is tested according to the received optical link test instructions. The test control center compares and analyzes the received test data with preset standard data, inputs the test data into a machine learning model for training and updating, and determines whether the optical link is faulty based on the comparative analysis results and prediction results. If a fault is present, the fault type and location are marked using a fault location algorithm. A test report is generated based on the comparative analysis results, fault judgment results, fault type, and fault location, and a fault solution is automatically generated based on the fault type. This method enables automated and intelligent testing of FTTR optical links, improving the efficiency and accuracy of fault diagnosis.

[0051] First, the machine learning model is used to dynamically generate test instructions. Combined with parameters such as the link length of the optical link, the optical module model, and the network topology structure, it can accurately test different optical link characteristics, thereby improving the accuracy and pertinence of the test.

[0052] Secondly, fault judgment based on comparative analysis results and the prediction results of the machine learning model can more comprehensively and accurately identify faults in optical links and reduce the misjudgment rate. The fault location algorithm can quickly and accurately mark the fault type and location, and can use the optical time domain reflection algorithm to accurately locate the optical fiber fault point.

[0053] Thirdly, by inputting test data into the machine learning model for training and updating, the model can continuously learn and adapt to changes in the optical link, improve the accuracy of fault prediction, formulate test instruction strategies based on the prediction results, and perform preventive testing before a fault occurs, discover potential problems in advance, and reduce the probability of failure.

[0054] Finally, when generating test instructions, the system optimizes them based on the performance and resource constraints of the FTTR equipment being tested, avoiding resource waste and improving equipment utilization. Automatically selecting the communication protocol through negotiation ensures stable communication between the test control center and the FTTR equipment, improving the system's compatibility and adaptability. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 The figure is a flow chart of the automated testing method for FTTR optical links of the present invention.

[0056] Figure 2 The figure is a flow chart of establishing a communication connection between a test control center and FTTR equipment according to the present invention. DETAILED DESCRIPTION

[0057] The following is combined with Figures 1 and 2 The present invention is described in further detail:

[0058] Example 1

[0059] See attached Figure 1 As shown, an automated testing method for an FTTR optical link includes the following steps:

[0060] S1: Collect the specified parameters of the FTTR optical link to be tested, including link length, optical module model, and network topology, and establish a communication connection between the test control center and the FTTR equipment;

[0061] S2: The test control center dynamically generates an optical link test instruction according to the specified parameters based on a built-in machine learning model, and sends the optical link test instruction to the FTTR device;

[0062] S3: The FTTR device controls the corresponding test module to test the optical link according to the received optical link test instruction, and feeds back the test data to the test control center;

[0063] S4: The test control center compares and analyzes the received test data with preset standard data, and inputs the test data into the machine learning model for training and updating;

[0064] S5: The test control center determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If there is a fault, the fault type and fault location are marked using a fault location algorithm.

[0065] S6: Generate a test report based on the comparison and analysis results, fault judgment results, fault type and fault location, and automatically generate a fault solution according to the fault type.

[0066] See attached Figure 2 As shown, the specific process of establishing the communication connection between the test control center and the FTTR device in step S1 is as follows:

[0067] S11: Start the test control center and send a device scanning signal to the network through the preset communication port. Use broadcast or multicast to detect the FTTR device to be tested. After receiving the scanning signal, the FTTR device to be tested returns a response data packet containing the device's unique identifier and a list of supported communication protocols.

[0068] S12: The test control center receives and parses the response data packet, and selects the communication protocol with the highest priority and supported by both parties from its own supported protocol library based on the communication protocol list fed back by the FTTR device to be tested, and sends a protocol negotiation request to the FTTR device. The two parties interactively negotiate the data packet to determine the parameters of the communication protocol version, data transmission format, and data compression method to be finally used, ensuring accurate transmission and parsing of data during the communication process;

[0069] S13: The test control center and the FTTR device can use pre-shared keys, digital certificates, or dynamic tokens for identity authentication. When using pre-shared keys, the test control center sends an encrypted authentication request to the FTTR device. The FTTR device uses the same pre-shared key to decrypt and verify the legitimacy of the request. If the verification is successful, it returns a successful authentication response. If using digital certificates, the two parties exchange certificates and verify them through a certificate authority to ensure the trustworthiness of the communicating parties.

[0070] S14: After authentication is successful, connection parameters are configured. The test control center negotiates with the FTTR device to determine the connection timeout, retransmission mechanism (including automatic repeat request (ARQ) parameter settings), and data cache size parameters. A port number is allocated for communication, and a stable data transmission channel is established. The test control center and the FTTR device establish a communication connection. Both parties continuously send heartbeat packets, which are periodic test data packets to monitor the connection status. If the test control center does not receive a heartbeat response from the FTTR device within the specified time, the connection is considered disconnected, triggering the reconnection mechanism and re-executing the device scanning, protocol negotiation, and security authentication processes until the communication connection is restored.

[0071] Example 2

[0072] On the basis of Example 1, the specific process of the test control center dynamically generating the optical link test instruction according to the specified parameters based on the built-in machine learning model in step S2 is as follows:

[0073] S21: Integrate the specified parameters to form a structured data set, and convert the data in the structured data set by label encoding;

[0074] S22: Feature extraction is performed on the converted structured data set to screen out key features that are strongly correlated with optical link failures to obtain strongly correlated features, thereby reducing data dimensions and improving model processing efficiency.

[0075] S23: Inputting the strongly correlated features into a built-in machine learning model, analyzing the data features layer by layer through multiple layers or nodes of the model, identifying patterns and regularities in the data, and predicting the type of optical link faults and the peak fault period under specified link lengths, optical module models, and environmental conditions;

[0076] S24: Based on the prediction results, the test control center formulates an optical link test instruction strategy. Based on the formulated test instruction strategy, specific optical link test instructions are generated, including optical power test instructions, bit error rate test instructions, and link connectivity test instructions. During the instruction generation process, the instructions are optimized based on the performance and resource limitations of the FTTR equipment to be tested. The execution order of the test instructions is set to avoid running multiple high-resource consumption tests simultaneously. The parameters in the test instructions, including the test time interval and test duration, are dynamically adjusted to ensure that the instructions can fully detect the optical link status without placing an excessive burden on the FTTR equipment.

[0077] S25: Before sending the generated optical link test command, a preliminary verification is performed. This is done by simulating the command's execution on the FTTR device to check the command's integrity, accuracy, and feasibility. If any issues are found during the verification process, including command format errors or parameters outside the device's supported range, the system returns to the command generation and optimization step for correction. Once the verification is successful, the optical link test command is sent to the FTTR device, and the command-related information is recorded for subsequent analysis and traceability of the test results.

[0078] In this embodiment, parameters such as link length (numeric), optical module model (text), and network topology (classification / graph structure) are obtained from a database or real-time interface. A two-dimensional table structure is constructed, with each row representing a test scenario and each column corresponding to a parameter. A dictionary mapping is established for the optical module model, and the network topology is converted into a categorical variable using one-hot encoding. Feature extraction includes basic statistical features such as the mean / variance of link length and the optical module usage time, the combined feature of link length × optical module loss coefficient, and the temporal feature of the periodicity of the test time. The correlation coefficient between the features and the fault type is calculated, and a feature importance matrix is ​​constructed. A multilayer perceptron is used with the following structure: an input layer containing 10 nodes → a hidden layer containing 20 nodes of ReLU → a hidden layer containing 15 nodes of Tanh → an output layer containing 3 nodes of Softmax. Fault type prediction is performed in combination with random forest and gradient boosting trees. The dataset is divided into 70% training set, 15% validation set, and 15% test set. The model is input for training, testing, and validation, and finally fault prediction is achieved. The prediction results include information such as the probability distribution of fault types and the peak fault period. Dynamically adjust test focus and test parameters based on fault prediction results, simulate the instruction processing logic of FTTR equipment, and verify and correct test parameters.

[0079] In step S4, the test control center compares and analyzes the received test data with the preset standard data, and inputs the test data into the machine learning model for training and updating. The specific process is as follows:

[0080] S41: The test control center removes duplicate data from the received test data and determines the uniqueness of the data through the timestamp and device identification. It checks the data integrity and uses adjacent value interpolation and multiple filling methods to fill in missing data when the missing ratio does not exceed the preset threshold. If the missing ratio exceeds the preset threshold, the data is marked and used with caution in subsequent analysis. The data is formatted and unified into a format that can be recognized by the machine learning model, such as converting the optical power value from a string type to a floating point type.

[0081] S42: Compare the pre-processed test data with the preset standard data. For the optical power value, determine whether it is within the standard power range. If it is out of range, calculate the deviation value and analyze the possible causes of the deviation in combination with factors such as link length and optical module model; for the number of bit errors, compare it with the preset bit error rate threshold. If the bit error rate is too high, mark the link as having a possible signal transmission problem; for the link connectivity status, compare it with the standard of the normal connectivity status to determine whether the link is interrupted or unstable. Through comparative analysis, generate preliminary fault judgment results, mark the optical links that may have faults and preliminary judgment information of their fault types;

[0082] S43: Perform feature extraction on the test data. For optical power data, extract features such as power fluctuation amplitude and power change trend; for bit error data, extract features such as bit error rate change rate and bit error distribution pattern. Utilize feature selection algorithms to screen out key features that are highly correlated with optical link faults, remove redundant and irrelevant features, reduce data dimensionality, and improve model training efficiency. Initial fault diagnosis results are used as label data and integrated with the extracted feature data to form a dataset for model training.

[0083] S44: Divide the integrated data set into a training set and a validation set, randomly divide them according to the specified ratio, use the training set to train the built-in machine learning model, and select appropriate training algorithms and hyperparameters. During the training process, cross-validation is used to avoid model overfitting, and model performance is optimized by adjusting hyperparameters. After training is completed, the model is evaluated using the validation set, and the model's accuracy, recall rate, and F1 value evaluation indicators are calculated. If the evaluation results do not meet the preset standards, the hyperparameters are readjusted or other training algorithms are selected, and training and evaluation are performed again until the model performance meets the requirements. The updated model is saved for subsequent test instruction generation and fault prediction.

[0084] Example 3

[0085] Based on Example 1 or Example 2, the test control center in step S5 determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If a fault exists, the specific process of marking the fault type and fault location using the fault location algorithm is as follows:

[0086] S51: The test control center integrates the comparative analysis results with the prediction results of the machine learning model. The comparative analysis results include the deviation of optical power, bit error rate, link connectivity indicators from standard data and preliminary fault judgment. The model prediction results are based on the prediction of the current optical link status after historical data training. The two types of results are integrated according to key information such as timestamps and optical link identifiers to form a unified judgment data set to ensure comprehensive information and accurate correlation.

[0087] S52: Perform preliminary fault diagnosis based on the fused data set. For optical power, if the comparative analysis results show that the power exceeds the standard range and the machine learning model also predicts that there is a risk of power anomaly, it is preliminarily determined that there is a fault in the optical power-related link; for bit error rate, when the actual bit error rate is higher than the threshold and the model predicts that the bit error rate is on an upward trend, the bit error problem is marked; for link connectivity, if the comparative analysis determines that the link is interrupted and the model predicts that the connection is unstable, it is determined that there is a link connectivity fault. After preliminary judgment, the optical link suspected of having a fault is screened out;

[0088] S53: For optical links initially identified as faulty, the fault type is further refined. The machine learning model's prediction results are used to identify the fault mode, combined with the specific deviations of each indicator during comparative analysis. If the optical power continues to decrease and the model predicts power attenuation due to optical module aging, the fault type is determined to be optical module aging, combined with information such as the optical module model and usage time. If the bit error rate fluctuates dramatically and the model suggests excessive optical fiber loss, the link length and environmental factors are considered to determine that the fault type is abnormal optical fiber loss.

[0089] S54: After determining the fault type, a fault location algorithm is used to mark the fault location. For optical power-related faults, if the fault is determined to be an optical module problem, the optical link topology is queried to locate the FTTR device node where the optical module is installed. If the fault is a fiber loss fault, the optical time domain reflectometry algorithm is used to calculate the location of the fiber fault point based on the optical signal reflection time and intensity. For link connectivity faults, the fault is checked step by step from the trunk link to the branch link. Based on the network topology and device connection relationships, the specific link segment or connection node where the fault occurred is determined through segmented testing.

[0090] S55: After completing the marking of the fault type and location, verify the results and retrieve relevant test data and historical records again to verify the rationality of the fault judgment and location. If there are contradictions or uncertainties, re-execute part of the judgment process, analyze the model prediction data again, supplement the specified test items, etc., to ensure that the fault type and location information are accurate, and finally form a complete fault marking result to provide a reliable basis for subsequent maintenance and processing.

[0091] In step S54, the specific process of calculating the location of the optical fiber fault point based on the reflection time and intensity of the optical signal using the optical time domain reflectometry algorithm is as follows:

[0092] S541: The test control center controls the optical transmitter module in the FTTR equipment to send a series of optical pulse signals to the optical fiber link under test. These optical pulse signals have a preset wavelength, pulse width, and power. The pulse width determines the duration of the optical pulse. A larger pulse width increases the detectable optical fiber distance, but the distance resolution decreases. The power affects the transmission distance of the optical signal in the optical fiber and the intensity of the reflected signal. The transmission power needs to be set according to the length and loss of the optical fiber link.

[0093] S542: When light pulses are transmitted through optical fibers, they generate Rayleigh scattering and Fresnel reflection due to the inhomogeneity of the fiber material and the locations of connectors, splices, and breakpoints within the fiber link. The optical receiver module in the FTTR equipment collects these reflected optical signals in real time and converts them into electrical signals. During the collection process, the optical receiver module must have sufficient sensitivity and sampling rate to accurately capture weak reflected signals and rapidly changing signal characteristics. The higher the sampling rate, the higher the temporal resolution of the reflected signals, and the more accurate the subsequent calculation of the fault location.

[0094] S543: Preprocess the collected electrical signals, including signal amplification, filtering, and analog-to-digital conversion. First, an amplifier is used to amplify the weak electrical signal, enhancing its strength for subsequent processing. Then, a filter is used to remove noise and interference from the signal. Common filters include low-pass filters and band-pass filters. The appropriate filtering method is selected based on the frequency characteristics of the signal. Finally, the analog electrical signal is converted into a digital signal for computer storage and analysis. The accuracy of the analog-to-digital conversion determines the accuracy of signal quantization, which affects the accuracy of subsequent fault point calculation.

[0095] S544: Analyze the pre-processed digital signal and plot a curve showing the optical power variation over time, i.e., the OTDR curve. On the OTDR curve, Rayleigh scattering of a normal fiber segment appears as a smooth slope change, while Fresnel reflection peaks appear at locations such as connectors, splices, and breakpoints. By analyzing the characteristics of the OTDR curve, parameters such as the position, height, and width of the reflection peak are identified. The position of the reflection peak corresponds to the round-trip transmission time of the optical signal, the height reflects the intensity of the reflected light signal, and the width is related to the characteristics of the reflection point.

[0096] S545: According to the propagation speed of the optical signal in the optical fiber and the round-trip time of the reflected optical signal, the propagation speed of the light in the optical fiber is about 2×10 8 m / s. The specific speed is related to the refractive index of the optical fiber. The distance from the fault point to the test end is calculated using the formula L = v × t / 2, where L is the distance to the fault point, v is the propagation speed of light in the optical fiber, and t is the time from the optical signal being emitted to being reflected back to the test end. By accurately measuring the time corresponding to the reflection peak position in the OTDR curve and combining it with the known optical fiber propagation speed, the location of the optical fiber fault point can be accurately calculated. The type of fault can be further determined based on the intensity and shape of the reflection peak. For example, a high-intensity, sharp reflection peak usually indicates a fiber break, while a relatively gentle reflection peak may indicate a bad connector or splice.

[0097] S546: Verify the calculated fault location, taking into account the effects of factors such as the actual fiber installation path and curvature on optical signal transmission time, and make corrections to the calculated results. If there are multiple reflection peaks or complex curve characteristics, compare historical OTDR data with the fiber link topology to eliminate misjudgments and ensure the accuracy of the fault location. The accurate fault location information is fed back to the test control center for generating a detailed fault report and repair instructions.

[0098] In step S6, a test report is generated based on the comparative analysis results, fault judgment results, fault type and fault location, and a specific process of automatically generating a fault solution according to the fault type is as follows:

[0099] S61: The test control center categorizes and organizes data such as comparative analysis results, fault diagnosis results, fault type, and fault location according to different dimensions. Comparative analysis data such as optical power, bit error rate, and link connectivity are classified as performance indicators; preliminary fault diagnosis and detailed fault type are classified as fault diagnosis; and fault location information is separately listed as fault location. At the same time, basic information such as test time, test environment, and FTTR equipment model is supplemented to complete the data set.

[0100] S62: Build the test report framework based on industry standards and actual application requirements. The report should include a cover page, table of contents, and overview. The overview should include sections such as the test purpose, scope, test results, fault details, and appendices. Ensure clear and logical content. The cover page should include the project name, test time, and generating unit. The test results should include comparative analysis data visualization charts and fault diagnosis conclusions. Fault details should include fault type, location description, and supporting materials such as OTDR curves. The appendices should include original test data and relevant technical parameters.

[0101] S63: Convert the classified data into intuitive visual charts. For indicators such as optical power and bit error rate in the comparative analysis results, use line charts to show their changing trends over time or link length; use bar charts to compare the differences between actual test data and standard data. For fault location information, combine it with the fiber link topology diagram to mark the specific location of the fault point, and use different colors or icons to distinguish the fault type, enhancing the readability and comprehensibility of the report.

[0102] S64: Fill the organized data and visual charts into the corresponding sections according to the established report structure.

[0103] In summary, the present invention provides an automated testing method for FTTR optical links.

Claims

1. An automated testing method for FTTR optical links, characterized in that: The following steps are involved: S1: Collect the specified parameters of the FTTR optical link to be tested, including link length, optical module model, and network topology, and establish a communication connection between the test control center and the FTTR equipment; S2: The test control center dynamically generates an optical link test instruction according to the specified parameters based on a built-in machine learning model, and sends the optical link test instruction to the FTTR device; S3: The FTTR device controls the corresponding test module to test the optical link according to the received optical link test instruction, and feeds back the test data to the test control center; S4: The test control center compares and analyzes the received test data with preset standard data, and inputs the test data into the machine learning model for training and updating; S5: The test control center determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If there is a fault, the fault type and fault location are marked using a fault location algorithm. S6: Generate a test report based on the comparison and analysis results, fault judgment results, fault type and fault location, and automatically generate a fault solution according to the fault type.

2. The automated testing method for an FTTR optical link according to claim 1, characterized in that: The specific process of establishing the communication connection between the test control center and the FTTR equipment in step S1 is as follows: S11: Start the test control center and send a device scanning signal to the network through the preset communication port. Use broadcast or multicast to detect the FTTR device to be tested. After receiving the scanning signal, the FTTR device to be tested returns a response data packet containing the device's unique identifier and a list of supported communication protocols. S12: The test control center receives and parses the response data packet, and selects a communication protocol with the highest priority and supported by both parties from its supported protocol library according to the communication protocol list fed back by the FTTR device to be tested, and sends a protocol negotiation request to the FTTR device; S13: The test control center performs identity authentication with the FTTR device. The test control center sends an encrypted authentication request to the FTTR device. The FTTR device decrypts the request using the pre-shared key and verifies the legitimacy of the request. If the authentication succeeds, an authentication success response is returned, and step S14 is executed. S14: Configure connection parameters. The test control center negotiates with the FTTR device to determine the connection timeout, retransmission mechanism, and data cache size, allocates a port number for communication, and establishes a data transmission channel.

3. The automated testing method for an FTTR optical link according to claim 1, wherein: The specific process of the test control center dynamically generating the optical link test instructions according to the specified parameters based on the built-in machine learning model in step S2 is as follows: S21: Integrate the specified parameters to form a structured data set, and convert the data in the structured data set by label encoding; S22: performing feature extraction on the converted structured data set, screening out key features that are strongly correlated with the optical link failure to obtain strongly correlated features; S23: Inputting the strong correlation features into a built-in machine learning model, using the model to identify patterns and regularities in the data, and predicting the type of optical link faults and the peak fault-incidence period under specified link lengths, optical module models, and environmental conditions; S24: The test control center formulates an optical link test instruction strategy based on the prediction results, generates specific optical link test instructions based on the formulated test instruction strategy, and optimizes the instructions in combination with the performance and resource limitations of the FTTR equipment to be tested. S25: Verify the optical link test instruction by simulating the execution process of the instruction on the FTTR device to be tested to check the integrity, accuracy and feasibility of the instruction.

4. The automated testing method for an FTTR optical link according to claim 1, wherein: In step S4, the test control center compares and analyzes the received test data with the preset standard data, and inputs the test data into the machine learning model for training and updating. The specific process is as follows: S41: The test control center removes duplicate data from the received test data, determines data uniqueness through timestamps and device identifiers, checks data integrity, completes missing data, and converts the data into a format recognizable by the machine learning model. S42: Compare and analyze the processed test data with preset standard data, generate a preliminary fault diagnosis result based on the comparison and analysis results, mark the optical link with possible faults and preliminary fault type diagnosis information; S43: Feature extraction is performed on the test data. For optical power data, features such as power fluctuation amplitude and power change trend are extracted. For bit error data, features such as bit error rate change rate and bit error distribution pattern are extracted. A specified feature selection algorithm is used to screen out key features that are highly correlated with optical link faults, remove redundant and irrelevant features, and reduce data dimensionality. The preliminary fault judgment results are used as label data and integrated with the extracted feature data to form a data set for model training. S44: Divide the integrated data set into a training set and a validation set, randomly divide them according to the specified ratio, use the training set to train the built-in machine learning model, and during the training process, use the cross-validation method to adjust the hyperparameters. Use the validation set to evaluate the model and calculate the model's specified evaluation indicators. If the evaluation result does not meet the preset standard, readjust the hyperparameters and train and evaluate again until the model performance meets the requirements. Save the updated model for subsequent test instruction generation and fault prediction.

5. The automated testing method for an FTTR optical link according to claim 4, characterized in that: In step S5, the test control center determines whether there is a fault in the optical link based on the comparative analysis results and the prediction results of the machine learning model. If there is a fault, the fault type and fault location are marked using the fault location algorithm. The specific process is as follows: S51: The test control center integrates the comparative analysis results with the prediction results. The comparative analysis results include the deviation of optical power, bit error rate, link connectivity indicators from standard data and preliminary fault judgment. The model prediction results are based on the prediction of the current optical link status after historical data training. The two types of results are integrated according to the specified key information to form a unified judgment data set. S52: Perform preliminary fault diagnosis based on the fused data set. For optical power, if the comparative analysis results show that the power exceeds the standard range and the machine learning model also predicts a power anomaly risk, a preliminary judgment is made that a fault exists in the optical power-related link. For bit error rate, if the actual bit error rate is higher than the threshold and the model predicts an upward trend in the bit error rate, a bit error problem is marked. For link connectivity, if the comparative analysis confirms a link interruption and the model predicts an unstable connection, a link connectivity fault is determined. After preliminary judgment, optical links suspected of having faults are screened out. S53: For optical links that are initially determined to have faults, the fault type is refined. Using the information about the fault mode in the prediction results and the specific deviations of each indicator during comparative analysis, if the optical power continues to decrease and the model predicts power attenuation due to optical module aging, the fault type is determined to be optical module aging, based on the optical module model and usage time. If the bit error rate fluctuates dramatically and the model indicates excessive optical fiber loss, the link length and environmental factors are considered to determine that the fault type is abnormal optical fiber loss. S54: After determining the fault type, a fault location algorithm is used to mark the fault location. For optical power-related faults, if it is determined to be an optical module problem, the optical link topology is queried to locate the FTTR device node where the optical module is installed. If the fault is fiber loss, the optical time domain reflectometry algorithm is used to calculate the location of the fiber fault point based on the optical signal reflection time and intensity. For link connectivity faults, the fault is checked step by step from the trunk link to the branch link. Based on the network topology and device connection relationships, the specific link section or connection node where the fault occurred is determined through segmented testing. S55: After completing the marking of the fault type and location, the results are verified, and the test data and historical records are retrieved again to verify the rationality of the fault judgment and location. If there are contradictions or uncertainties, part of the judgment process is re-executed to ensure that the fault type and location information are accurate and form a complete fault marking result.

6. The automated testing method for an FTTR optical link according to claim 5, characterized in that: In step S54, the specific process of calculating the location of the optical fiber fault point based on the reflection time and intensity of the optical signal using the optical time domain reflectometry algorithm is as follows: S541: The test control center controls the optical transmitter module in the FTTR device to send a series of optical pulse signals to the optical fiber link under test. The optical pulse signals have a preset wavelength, pulse width, and power. S542: When the optical pulse is transmitted in the optical fiber, the optical receiving module in the FTTR device collects the reflected optical signal in real time and converts the optical signal into an electrical signal; S543: Preprocessing the collected electrical signal: using an amplifier to amplify the weak electrical signal to enhance the signal strength, using a filter to remove noise and interference in the signal, and converting the analog electrical signal into a digital signal; S544: Analyze the pre-processed digital signal and plot a curve of optical power versus time, i.e., an OTDR curve. By analyzing the characteristics of the OTDR curve, the position, height, and width of the reflection peak are identified. The position of the reflection peak corresponds to the round-trip transmission time of the optical signal, the height reflects the intensity of the reflected optical signal, and the width reflects the characteristics of the reflection point. S545: Based on the propagation speed of the optical signal in the optical fiber and the round-trip time of the reflected optical signal, the distance from the fault point to the test end is calculated using the formula L = v × t / 2, where L is the distance to the fault point, v is the propagation speed of light in the optical fiber, and t is the time from the optical signal being transmitted to being reflected back to the test end. By measuring the time corresponding to the reflection peak position in the OTDR curve and combining it with the known optical fiber propagation speed, the location of the optical fiber fault point is calculated. The type of fault is determined based on the intensity and shape of the reflection peak. S546: Verify the calculated fault point location and correct the calculation result based on the actual fiber optic laying path and the impact of the bending degree on the optical signal transmission time. If there are multiple reflection peaks or complex curve characteristics, eliminate misjudgments by comparing historical OTDR data with the topology information of the fiber optic link, and feedback the fault point location information to the test control center.

7. The automated testing method for an FTTR optical link according to claim 1, characterized in that: In step S6, a test report is generated based on the comparative analysis results, fault judgment results, fault type and fault location, and a specific process of automatically generating a fault solution according to the fault type is as follows: S61: The test control center will compare the analysis results, fault judgment results, fault type and fault location data; S62: Build the framework structure of the test report according to industry standards and actual application requirements; S63: Convert classified and organized data into intuitive visual charts; S64: Fill the organized data and visual charts into the corresponding sections according to the established report structure.

Images