Optical module production monitoring method and system based on artificial intelligence, and storage medium

Abstract

The invention discloses an optical module production monitoring method and system based on artificial intelligence, and a storage medium, which are used for improving the product yield of optical modules in the production process. The method comprises the following steps: acquiring optical path parameters and temperature change data of a to-be-detected optical module; calculating the fluctuation similarity between the time sequences corresponding to the light path parameters and the temperature change data in the same time period, and determining the parameter drift trend of the light path parameters according to the fluctuation similarity; building a temperature field simulation model based on a preset temperature range, and calculating parameter drift distribution in the preset temperature range in combination with the parameter drift trend; establishing a thermal compensation model based on a Gaussian process regression algorithm and parameter drift distribution, and outputting a corrected value of the optical path parameter; adjusting light path alignment according to the correction value, and obtaining real-time coupling efficiency; judging whether the real-time coupling efficiency meets a preset requirement or not; and if not, updating the thermal compensation model until the real-time coupling efficiency meets the preset requirement.

Description

Technical Field

[0001] The present application relates to the technical field of optical module detection, and particularly to an optical module production monitoring method, system and storage medium based on artificial intelligence. Background Art

[0002] An optical module is an optoelectronic device for performing optoelectronic conversion and electro-optic conversion, and is one of the core components of a fiber optic communication system. Optical modules have a wide range of applications in fields such as 5G, cloud computing, and the Internet of Things. Their performance directly affects the efficiency and reliability of data transmission. Therefore, the research on the performance of optical modules has crucial strategic significance. With the improvement of communication rates and the increasing demand for device miniaturization, optical modules face increasingly important challenges in production processes, especially in terms of performance consistency and environmental adaptability. Any small deviation may lead to the overall failure of the system. Therefore, how to ensure the high quality and high stability of optical modules during the production process has become a key issue in the industry development.

[0003] Currently, during the production process of optical modules, static detection and manual adjustment are usually used to monitor the production of optical modules to ensure the production quality of optical modules. However, when facing complex environmental changes, the change in environmental temperature will cause the optical path parameters inside the optical module to drift. The existing production monitoring methods are difficult to reflect the influence of environmental temperature changes on the optical path parameters in real time, resulting in the inability to adjust in real time during optical path alignment, leading to deviations, and further affecting the coupling efficiency and signal transmission quality of the optical module, thus reducing the product yield of the optical module. Summary of the Invention

[0004] The present application provides an optical module production monitoring method, system and storage medium based on artificial intelligence, which is used to improve the product yield of optical modules during the production process.

[0005] The first aspect of the present application provides an optical module production monitoring method based on artificial intelligence, including:

[0006] Obtain the optical path parameters and temperature change data of the optical module to be detected;

[0007] Calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters according to the fluctuation similarity;

[0008] Build a temperature field simulation model based on a preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend;

[0009] Establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output the correction value of the optical path parameters;

[0010] Adjust the optical path alignment according to the correction value, and obtain the real-time coupling efficiency;

[0011] Judge whether the real-time coupling efficiency meets the preset requirements;

[0012] If not, update the thermal compensation model until the real-time coupling efficiency meets the preset requirements.

[0013] Optionally, calculating the fluctuation similarity between the optical path parameters and the time series corresponding to the temperature change data within the same time period, and determining the parameter drift trend of the optical path parameters according to the fluctuation similarity includes:

[0014] Use the dynamic time warping algorithm to calculate the fluctuation similarity between the optical path parameters and the time series corresponding to the temperature change data within the same time period;

[0015] Judge whether the fluctuation similarity is less than the similarity threshold;

[0016] If so, it is determined that there is a parameter drift in the optical path parameters during the time period;

[0017] Input the optical path parameters, the fluctuation similarity, and the parameter drift determination result into a long short-term memory network for training to obtain a parameter drift trend prediction model to output the parameter drift trend.

[0018] Optionally, the using the dynamic time warping algorithm to calculate the fluctuation similarity between the optical path parameters and the time series corresponding to the temperature change data within the same time period includes:

[0019] Intercept the time series corresponding to the optical path parameters and the temperature change data within the same time period;

[0020] Calculate the distance between each data point of the two time series;

[0021] Determine the path with the minimum cumulative distance;

[0022] Determine the cumulative distance of the path as the fluctuation similarity between the two time series.

[0023] Optionally, building a temperature field simulation model within a preset temperature range, and calculating the parameter drift distribution within the preset temperature range in combination with the parameter drift trend includes:

[0024] Use the environmental simulation algorithm to build a temperature field simulation model within a preset temperature range;

[0025] Use Monte Carlo simulation to simulate the random change values of the optical path parameters under different temperature changes, and generate offset data in combination with the Gaussian distribution;

[0026] Statistically analyze the offset data under different temperature changes in combination with the parameter drift trend to obtain a parameter drift distribution.

[0027] Optionally, establishing a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and outputting a correction value of the optical path parameter, including:

[0028] Extract feature points from the parameter drift distribution, where the feature points include inflection point features and extreme point features;

[0029] Use the Gaussian process regression algorithm to analyze and process the feature points to obtain the mapping relationship between temperature change and optical path parameters;

[0030] Establish a thermal compensation model according to the mapping relationship;

[0031] Obtain a set of compensation coefficients through the thermal compensation model;

[0032] When the offset of the optical path parameter within the temperature range in the set of compensation coefficients is greater than a preset offset threshold, use an adaptive control algorithm to calculate the correction value of the optical path parameter.

[0033] Optionally, adjusting the optical path alignment according to the correction value and obtaining the real-time coupling efficiency, including:

[0034] Generate a device drive instruction based on the correction value;

[0035] Send the device drive instruction to the optical path alignment device to instruct the optical path alignment device to adjust the optical path alignment action;

[0036] Obtain the optical path parameter after the optical path alignment is completed;

[0037] Optimize the optical path parameter based on the clustering algorithm and the interpolation method;

[0038] Obtain the real-time coupling efficiency from the optimized optical path parameter.

[0039] Optionally, updating the thermal compensation model until the real-time coupling efficiency meets the preset requirements, including:

[0040] Use the isolation forest algorithm to detect abnormal data in the optical path parameter;

[0041] Analyze the deviation source of the abnormal data, and optimize the optical path parameter according to the deviation analysis result and in combination with the genetic algorithm;

[0042] Iteratively update the compensation coefficients of the thermal compensation model through the optimized optical path parameter until the updated real-time coupling efficiency meets the preset requirements.

[0043] The second aspect of the present application provides an optical module production monitoring system based on artificial intelligence, including:

[0044] An acquisition unit for acquiring optical path parameters and temperature change data of an optical module to be detected;

[0045] A calculation unit for calculating the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determining the parameter drift trend of the optical path parameters according to the fluctuation similarity;

[0046] A building unit for building a temperature field simulation model within a preset temperature range, and calculating the parameter drift distribution within the preset temperature range in combination with the parameter drift trend;

[0047] A establishing unit for establishing a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and outputting a correction value of the optical path parameters;

[0048] An adjustment unit for adjusting the optical path alignment according to the correction value and obtaining the real-time coupling efficiency;

[0049] A judgment unit for judging whether the real-time coupling efficiency meets a preset requirement;

[0050] An updating unit for updating the thermal compensation model until the real-time coupling efficiency meets the preset requirement.

[0051] The third aspect of the present application provides an electronic device, including:

[0052] A central processing unit, a memory, an input / output interface, a wired or wireless network interface, and a power supply;

[0053] The memory is a transient storage memory or a persistent storage memory;

[0054] The central processing unit is configured to communicate with the memory and execute the instruction operations in the memory to execute any one of the methods in the first aspect and the optional manners of the first aspect.

[0055] The fourth aspect of the present application provides a computer-readable storage medium, including instructions, which when running on a computer, cause the computer to execute any one of the methods in the first aspect and the optional manners of the first aspect.

[0056] It can be seen from the above technical solutions that the present application has the following effects:

[0057] First, obtain the optical path parameters and temperature change data of the optical module to be detected; then calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters based on the fluctuation similarity; next, build a temperature field simulation model within a preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend; further, establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output the corrected value of the optical path parameters; further adjust the optical path alignment according to the corrected value, and obtain the real-time coupling efficiency; finally, determine whether the real-time coupling efficiency meets the preset requirements; if not, update the thermal compensation model until the real-time coupling efficiency meets the preset requirements. In this way, by building a temperature field simulation model, the application scenarios of different environments can be simulated, and the influence of temperature changes on the optical path parameters can be mapped by the parameter drift distribution calculated by the temperature field simulation model. Then, a thermal compensation model that can output the corrected value of the optical path parameters is established according to the parameter drift distribution to adjust the optical path alignment in real time. Finally, by monitoring the real-time coupling efficiency to update the thermal compensation model in real time, the dynamic adjustment of the real-time coupling efficiency is realized, so as to improve the coupling efficiency and stability of the optical module, thereby improving the product yield of the optical module in the production process. Brief Description of the Drawings

[0058] Figure 1 is a schematic diagram of an embodiment of a method for monitoring the production of an optical module based on artificial intelligence in the present application;

[0059] Figure 2-1 、 Figure 2-2 and Figure 2-3 is a schematic diagram of another embodiment of a method for monitoring the production of an optical module based on artificial intelligence in the present application;

[0060] Figure 3 is a schematic diagram of an embodiment of a system for monitoring the production of an optical module based on artificial intelligence in the present application;

[0061] Figure 4 is a schematic diagram of an embodiment of an electronic device in the present application. Detailed Description of the Embodiments

[0062] The present application provides a method, system and storage medium for monitoring the production of an optical module based on artificial intelligence, which is used to improve the product yield of the optical module in the production process.

[0063] The method for monitoring the production of an optical module based on artificial intelligence described in the present application is implemented by being executed on a terminal, a system, a server or other electronic devices with logical analysis and processing capabilities. The embodiments of the present application are described by taking the application on a terminal as an example. Please refer to Figure 1 As shown, an embodiment of the method for monitoring the production of an optical module based on artificial intelligence in the present application includes:

[0064] 101. Obtain the optical path parameters and temperature change data of the optical module to be detected;

[0065] In this embodiment, first, the shell temperature of the optical module to be detected is collected by a temperature sensor to obtain the temperature change data of the optical module to be detected; data such as the insertion loss, polarization-dependent loss, and wavelength shift of the optical module to be detected are collected by a optical power meter, a polarization controller, and a spectral analyzer respectively to obtain the optical path parameters of the optical module to be detected. Devices such as the temperature sensor, optical power meter, polarization controller, and spectral analyzer are communicatively connected to a terminal to transmit the collected temperature change data and optical path parameters to the terminal. In another implementable manner, a sensor array composed of multiple devices such as temperature sensors, optical power meters, polarization controllers, and spectral analyzers can be deployed on the production line of the optical module to simultaneously collect data from multiple optical modules to be detected on the production line, so as to improve the collection efficiency of the optical path parameters and temperature change data.

[0066] 102. Calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters according to the fluctuation similarity;

[0067] In this embodiment, after the terminal obtains the optical path parameters and temperature change data of the optical module to be detected, it extracts the optical path parameters and temperature change data within the same time period and sets a certain time step to generate two time series. For example: the time period is set to 1 - 4 s, the time step is set to 1 s, the time series of the optical path parameters within this time period is [a, b, c, a], and the time series of the temperature change data within this time period is [20 °C, 21 °C, 22 °C, 20 °C]. Then, methods such as Pearson correlation coefficient, dynamic time warping, cross-correlation analysis, or wavelet coherence analysis are used to calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data. When the fluctuation similarity does not meet the preset conditions, it indicates that the corresponding optical path parameter has a parameter drift, and then a neural network model is used to predict the trend of the fluctuation similarity to determine the parameter drift trend of the optical path parameters.

[0068] 103. Build a temperature field simulation model within a preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend;

[0069] During the production process of the optical module, when the temperature of the environment where the optical module is located changes, the coefficient of thermal expansion of the material of the optical module housing will change with the temperature, which will in turn cause the optical path alignment to drift. Therefore, simulation technology can be used to simulate application scenarios with multiple different environmental temperatures to calculate the parameter drift corresponding to different application scenarios, so as to improve the adaptability of the optical module to environmental changes. Among them, a virtual environment algorithm can be used to generate a virtual scenario to build a temperature field simulation model within a preset temperature range, calculate the corresponding relationship between the optical path parameters and temperature change data under different scenarios through this temperature field simulation model, and finally determine the parameter drift distribution of different scenarios in combination with the parameter drift trend.

[0070] 104. Establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output the corrected value of the optical path parameters;

[0071] In this embodiment, the Gaussian process regression algorithm is a set of infinitely many random variables defined on a continuous domain, and any finite number of random variables follow a joint Gaussian distribution. It is a non-parametric Bayesian regression method and has excellent performance in machine learning and small-sample regression problems. After obtaining the parameter drift distribution of different scenarios, use the parameter drift distribution as a training sample and establish a thermal compensation model using the Gaussian regression algorithm. Calculate the corrected value of the optical path parameters through the compensation coefficient output by this thermal compensation model to dynamically adjust the optical path parameters under different scenarios.

[0072] 105. Adjust the optical path alignment according to the corrected value and obtain the real-time coupling efficiency;

[0073] In this embodiment, after obtaining the corrected value of the optical path parameters, send an adjustment instruction to the optical path alignment device according to this corrected value. After the optical path alignment device receives the adjustment instruction and executes it, the terminal obtains the latest optical path parameters and calculates the real-time coupling efficiency according to the latest optical path parameters.

[0074] 106. Determine whether the real-time coupling efficiency meets the preset requirements. If not, execute step 107;

[0075] 107. Update the thermal compensation model until the real-time coupling efficiency meets the preset requirements.

[0076] After calculating the real-time coupling efficiency of the optical module to be detected, compare it with the production quality standard. If the real-time coupling efficiency is greater than or equal to the preset coupling efficiency threshold, it means that the optical module to be detected meets the production requirements, and at this time, this production monitoring process can be ended; if the real-time coupling efficiency is less than the preset coupling efficiency threshold, it means that the optical module to be detected does not meet the production requirements. At this time, the thermal compensation model needs to be updated, the optical path parameters need to be corrected continuously, and the real-time coupling efficiency is recalculated according to the corrected optical path parameters. This cycle continues until the latest real-time coupling efficiency is greater than or equal to the preset coupling efficiency threshold, realizing the closed-loop control of the real-time coupling efficiency.

[0077] In this embodiment, first, obtain the optical path parameters and temperature change data of the optical module to be detected; then calculate the fluctuation similarity between the time series corresponding to the optical path parameters and temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters according to the fluctuation similarity; then build a temperature field simulation model within the preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend; furthermore, establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output the correction value of the optical path parameters; further adjust the optical path alignment according to the correction value, and obtain the real-time coupling efficiency; finally, determine whether the real-time coupling efficiency meets the preset requirements; if not, update the thermal compensation model until the real-time coupling efficiency meets the preset requirements. In this way, by building a temperature field simulation model, the application scenarios of different environments can be simulated, and the influence of temperature change on the optical path parameters can be mapped by the parameter drift distribution calculated by the temperature field simulation model. Then, a thermal compensation model that can output the correction value of the optical path parameters is established according to the parameter drift distribution to adjust the optical path alignment in real time. Finally, by monitoring the real-time coupling efficiency to update the thermal compensation model in real time, the dynamic adjustment of the real-time coupling efficiency is realized, so as to improve the coupling efficiency and stability of the optical module, thereby improving the product yield of the optical module in the production process.

[0078] Please refer to Figures 2-1 to 2-3 As shown, another embodiment of the optical module production monitoring method based on artificial intelligence in this application includes:

[0079] 201. Obtain the optical path parameters and temperature change data of the optical module to be detected;

[0080] Step 201 in this embodiment is the same as step 101 in the foregoing Figure 1 shown embodiment, and will not be elaborated here.

[0081] 202. Use the dynamic time warping algorithm to calculate the fluctuation similarity between the time series corresponding to the optical path parameters and temperature change data within the same time period;

[0082] 203. Determine whether the fluctuation similarity is less than the similarity threshold. If so, execute step 204;

[0083] 204. Determine that there is a parameter drift in the optical path parameters within the time period;

[0084] 205. Input the optical path parameters, fluctuation similarity, and parameter drift determination result into a long short-term memory network for training to obtain a parameter drift trend prediction model to output the parameter drift trend;

[0085] Optionally, in this embodiment, the dynamic time warping algorithm can be used to calculate the fluctuation similarity between two different time series of optical path parameters and temperature change data. Specifically, first intercept the time series corresponding to the optical path parameters and temperature change data within the same time period. The optical path parameter time series can be set as X = x1, x2, …, x n ; The temperature change data time series can be set as Y = y1, y2,..., y n . Then calculate the distance between each data point of the two time series to construct an n×n distance matrix D, where D(i, j) = (x i - y j ) 2 ; Then construct the cumulative distance matrix C, where,

[0086] C(1, 1) = D(1, 1)

[0087] C(i, 1) = D(i, 1) + C(i - 1, 1)

[0088] C(1, j) = D(1, j) + C(1, j - 1)

[0089] C(i, j) = D(i, j) + min{C(i - 1, j), C(i, j - 1), C(i - 1, j - 1)}

[0090] Finally, determine the path with the minimum cumulative distance and determine the cumulative distance of the path as the fluctuation similarity between the two time series.

[0091] It should be noted that before calculating the fluctuation similarity between two different time series of optical path parameters and temperature change data, the optical path parameters and temperature change data can be preprocessed to reduce the noise interference of the optical path parameters and temperature change data and improve the data accuracy and stability. Specifically, first, perform low-pass filtering on the collected optical path parameters, use a Butterworth filter to remove high-frequency interference signals, and retain the effective optical path parameters. Subsequently, extract the time-frequency characteristics of the optical path parameters through short-time Fourier transform to capture the periodic law of parameter changes. For the temperature change data, perform multi-scale analysis using wavelet transform to identify the pattern of temperature gradient changes. Combine the corresponding relationship between the optical path parameters and temperature change data to construct a support vector machine regression model to predict the trend of optical path parameters with temperature changes.

[0092] Compare the calculated fluctuation similarity with the similarity threshold. If the fluctuation similarity is less than the similarity threshold, it indicates that the fluctuation amplitude of the optical path parameters has become abnormal, and it can be determined that the optical path parameters have undergone parameter drift. At this time, after further performing dimensionality reduction processing using the principal component analysis method, the optical path parameters, fluctuation similarity, and parameter drift determination results can be used as training samples and input into a long short-term memory network for model training to obtain a parameter drift trend prediction model that can predict the parameter drift trend.

[0093] 206. Use the environmental simulation algorithm to build a temperature field simulation model within a preset temperature range;

[0094] 207. Use Monte Carlo simulation to simulate the random change values of the optical path parameters under different temperature changes, and generate offset data in combination with the Gaussian distribution;

[0095] 208. Combine the parameter drift trend to perform statistical analysis on the offset data under different temperature changes to obtain the parameter drift distribution;

[0096] Optionally, in this embodiment, the environmental simulation algorithm can be used to generate a virtual scene, construct a dynamic sequence of temperature changes, and obtain the initial offset data of the optical path parameters. Through numerical simulation processing of the initial offset data, calculate the offset value of the optical path parameters under temperature changes to determine the preliminary form of the drift distribution. In the environmental simulation algorithm, use the finite element analysis method to build a temperature field simulation model, where the preset temperature range can be set from -40°C to 85°C, the temperature step is 5°C, and the temperature distribution is calculated through the heat conduction equation. Then use the Monte Carlo simulation method to simulate the random changes of the optical path parameters at different temperatures, generate offset data in combination with the Gaussian distribution. Perform statistical analysis on the offset through kernel density estimation, and then combine the parameter drift trend prediction to generate the parameter drift distributions of multiple different scenarios, providing a reliable basis for subsequent thermal compensation calculations.

[0097] 209. Extract feature points from the parameter drift distribution, where the feature points include inflection point features and extreme point features;

[0098] 210. Use the Gaussian process regression algorithm to analyze and process the feature points to obtain the mapping relationship between temperature change and optical path parameters;

[0099] 211. Establish a thermal compensation model according to the mapping relationship;

[0100] 212. Obtain a set of compensation coefficients through the thermal compensation model;

[0101] 213. When the offset of the optical path parameters within a temperature range in the set of compensation coefficients is greater than a preset offset threshold, use an adaptive control algorithm to calculate the correction value of the optical path parameters;

[0102] Optionally, in this embodiment, first extract key feature points from the parameter drift distribution. The feature points can include inflection point features and extreme point features. Among them, the inflection point feature represents the feature of the temperature sensitivity mutation site, and the second derivative zero-crossing detection algorithm can be used to extract the feature points; the extreme point feature represents the feature of the maximum or minimum drift position, and the first derivative zero-crossing detection algorithm can be used to extract the feature points. It can be understood that other features such as the midpoint of the linear region and the mutation point can also be extracted from the parameter drift distribution. The midpoint of the linear region represents the optimal compensation working point, and the piecewise linear fitting algorithm can be used to extract the feature points; the mutation point represents the potential fault region, and the sliding window difference detection algorithm can be used to extract the feature points. After obtaining the required feature points, use the Gaussian process regression algorithm to analyze the mapping relationship between temperature change and optical path parameters to establish a thermal compensation model, and then optimize the mapping relationship by the least squares method to obtain a set of compensation coefficients. Further, use the support vector machine algorithm to classify the compensation coefficients, and combine the decision tree algorithm to verify the classification results to ensure the accuracy of the compensation coefficients. Then use the gradient boosting algorithm to optimize the compensation coefficients, and combine the random forest algorithm to evaluate the optimization results to verify the rationality of the optimization.

[0103] Next, determine whether there is an offset value of the optical path parameters in a certain temperature range in the set of compensation coefficients greater than the preset offset value. If so, trigger the dynamic adjustment algorithm for real-time correction. First, perform state estimation on the current optical path parameters based on the Kalman filter algorithm, and update the optical path deviation model in combination with the recursive least squares method. Subsequently, use the adaptive control algorithm to calculate the correction value of the optical path alignment and generate a set of device drive instructions. To optimize the adjustment effect, introduce the particle swarm optimization algorithm to globally search for the correction value to ensure the accuracy of the adjustment instructions.

[0104] 214. Generate device drive instructions based on the correction value;

[0105] 215. Send device driver instructions to the optical path alignment device to instruct the optical path alignment device to adjust the optical path alignment operation;

[0106] 216. Obtain the optical path parameters after the optical path alignment is completed;

[0107] 217. Optimize the optical path parameters based on the clustering algorithm and the interpolation method;

[0108] 218. Obtain the real-time coupling efficiency from the optimized optical path parameters;

[0109] Optionally, in this embodiment, device driver instructions are generated according to the correction value. After the optical path alignment device responds to the device driver instructions, a dynamic adjustment operation is performed to obtain the adjusted operating state data of the optical module to be detected. The optical path parameters are obtained from the adjusted operating state data of the optical module to be detected. Feature values are extracted through real-time data processing to determine the feature value distribution. The clustering algorithm is used to divide the data interval for the feature value distribution to obtain the classified set of optical path parameters. If there is a deviation between the classified set of optical path parameters and the preset parameter threshold, the feature value distribution is adjusted by the interpolation method to obtain the optimized set of optical path parameters, and the real-time coupling efficiency is calculated according to the optimized set of optical path parameters.

[0110] Among them, after the piezoelectric ceramic actuator that drives the optical path alignment device receives the device driver instructions, it adjusts the lens pitch angle and lateral displacement in real time through the PID control algorithm. And the differential evolution algorithm is used to optimize the execution trajectory, and the coupling efficiency data is collected in combination with the optical power meter.

[0111] 219. Determine whether the real-time coupling efficiency meets the preset requirements. If so, execute step 220;

[0112] 220. Use the isolation forest algorithm to detect abnormal data in the optical path parameters;

[0113] 221. Analyze the deviation source of the abnormal data, and optimize the optical path parameters according to the deviation analysis results and in combination with the genetic algorithm;

[0114] 222. Iteratively update the compensation coefficient of the thermal compensation model through the optimized optical path parameters until the updated real-time coupling efficiency meets the preset requirements.

[0115] Optionally, in this embodiment, when it is detected that the real-time coupling efficiency does not meet the preset requirements, that is, the real-time coupling efficiency is less than the preset coupling efficiency threshold, the isolation forest algorithm is used to identify abnormal data from the current optical path parameters, and the principal component analysis is combined to extract the features of the abnormal data. For the extracted features, a linear regression model is used to analyze the deviation source, generate a deviation distribution map, and the parameter set is optimized by the genetic algorithm to generate an optimized parameter value. The current optical path parameters are optimized using the optimized parameter value. The compensation coefficient in the thermal compensation model is updated with the optimized optical path parameters to obtain the updated compensation coefficient. According to the updated compensation coefficient, the model parameters are adjusted by iterative training to complete the iterative update of the thermal compensation model. The latest correction value output by the updated thermal compensation model is obtained, and the optical path alignment adjustment operation is re-performed according to the latest correction value, and the latest real-time coupling efficiency is obtained, and so on in a loop until the real-time coupling efficiency meets the preset requirements, so as to realize the dynamic adjustment and closed-loop control of the real-time coupling efficiency.

[0116] Please refer to Figure 3 As shown, an embodiment of the optical module production monitoring system based on artificial intelligence in the present application includes:

[0117] An acquisition unit 301, configured to acquire the optical path parameters and temperature change data of the optical module to be detected;

[0118] A calculation unit 302, configured to calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters according to the fluctuation similarity;

[0119] A building unit 303, configured to build a temperature field simulation model within a preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend;

[0120] An establishment unit 304, configured to establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output a correction value of the optical path parameters;

[0121] An adjustment unit 305, configured to adjust the optical path alignment according to the correction value and obtain the real-time coupling efficiency;

[0122] A judgment unit 306, configured to judge whether the real-time coupling efficiency meets the preset requirements;

[0123] An update unit 307, configured to update the thermal compensation model until the real-time coupling efficiency meets the preset requirements.

[0124] In this embodiment, the acquisition unit 301 acquires the optical path parameters and temperature change data of the optical module to be detected; the calculation unit 302 calculates the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determines the parameter drift trend of the optical path parameters according to the fluctuation similarity; the building unit 303 builds a temperature field simulation model within a preset temperature range, and calculates the parameter drift distribution within the preset temperature range in combination with the parameter drift trend; the establishment unit 304 establishes a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and outputs the correction value of the optical path parameters; the adjustment unit 305 adjusts the optical path alignment according to the correction value, and obtains the real-time coupling efficiency; the judgment unit 306 judges whether the real-time coupling efficiency meets the preset requirements; the update unit 307 updates the thermal compensation model until the real-time coupling efficiency meets the preset requirements. In this way, by building a temperature field simulation model, the application scenarios of different environments can be simulated, and the influence of temperature change on the optical path parameters can be mapped by the parameter drift distribution calculated by the temperature field simulation model. Then, a thermal compensation model that can output the correction value of the optical path parameters is established according to the parameter drift distribution to adjust the optical path alignment in real time. Finally, by monitoring the real-time coupling efficiency to update the thermal compensation model in real time, the dynamic adjustment of the real-time coupling efficiency is realized, so as to improve the coupling efficiency and stability of the optical module, thereby improving the product yield of the optical module in the production process.

[0125] Please refer to Figure 4 As shown, an embodiment of the electronic device in the present application includes:

[0126] A central processing unit 402, a memory 401, an input / output interface 403, a wired or wireless network interface 404, and a power supply 405;

[0127] The memory 401 is a transient storage memory or a persistent storage memory;

[0128] The central processing unit 402 is configured to communicate with the memory 401 and execute the instruction operations in the memory 401 to execute the steps in the foregoing Figure 1 , Figure 2-1 , Figure 2-2 and Figure 2-3 shown embodiments.

[0129] The present application provides a computer-readable storage medium, including instructions, which when run on a computer, cause the computer to execute the steps in the foregoing Figure 1 , Figure 2-1 , Figure 2-2 and Figure 2-3 shown embodiments.

[0130] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0131] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0132] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0133] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of software functional units.

[0134] If the above-mentioned integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of this 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 for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, read-only memory), random access memories (RAM, random access memory), magnetic disks, or optical discs that can store program codes.

Claims

1. An optical module production monitoring method based on artificial intelligence, characterized in that, Including: Obtain the optical path parameters and temperature change data of the optical module to be detected; Calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determine the parameter drift trend of the optical path parameters according to the fluctuation similarity; Build a temperature field simulation model within a preset temperature range, and calculate the parameter drift distribution within the preset temperature range in combination with the parameter drift trend; Establish a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and output the corrected value of the optical path parameters; Adjust the optical path alignment according to the corrected value, and obtain the real-time coupling efficiency; Judge whether the real-time coupling efficiency meets the preset requirements; If not, update the thermal compensation model until the real-time coupling efficiency meets the preset requirements.

2. The optical module production monitoring method according to claim 1, wherein The calculating the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determining the parameter drift trend of the optical path parameters according to the fluctuation similarity includes: Use the dynamic time warping algorithm to calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period; Judge whether the fluctuation similarity is less than the similarity threshold; If so, determine that there is a parameter drift in the optical path parameters during the time period; Input the optical path parameters, the fluctuation similarity, and the parameter drift determination result into a long short-term memory network for training to obtain a parameter drift trend prediction model, so as to output the parameter drift trend.

3. The optical module production monitoring method according to claim 2, wherein, The using the dynamic time warping algorithm to calculate the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period includes: Intercept the time series corresponding to the optical path parameters and the temperature change data within the same time period; Calculate the distance between each data point of the two time series; Determine the path with the minimum cumulative distance; Determine that the cumulative distance of the path is the fluctuation similarity between the two time series.

4. The optical module production monitoring method according to claim 1, wherein The building a temperature field simulation model within a preset temperature range, and calculating the parameter drift distribution within the preset temperature range in combination with the parameter drift trend includes: Use the environmental simulation algorithm to build a temperature field simulation model within a preset temperature range; Use Monte Carlo simulation to simulate the random change values of the optical path parameters under different temperature changes, and generate offset data in combination with the Gaussian distribution; Conduct statistical analysis on the offset data under different temperature changes in combination with the parameter drift trend to obtain the parameter drift distribution.

5. The optical module production monitoring method according to claim 1, characterized in that The establishing a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and outputting the corrected value of the optical path parameters includes: Extract feature points from the parameter drift distribution, and the feature points include inflection point features and extreme point features; Use the Gaussian process regression algorithm to analyze and process the feature points to obtain the mapping relationship between temperature change and optical path parameters; Establish a thermal compensation model according to the mapping relationship; Obtain a set of compensation coefficients through the thermal compensation model; When the offset of the optical path parameters within the temperature range in the compensation coefficient set is greater than the preset offset threshold, an adaptive control algorithm is used to calculate the correction value of the optical path parameters.

6. The optical module production monitoring method according to claim 1, wherein The adjusting the optical path alignment according to the correction value and obtaining the real-time coupling efficiency includes: Generating a device drive instruction based on the correction value; Sending the device drive instruction to the optical path alignment device to instruct the optical path alignment device to adjust the optical path alignment action; Obtaining the optical path parameters after the optical path alignment is completed; Optimizing the optical path parameters based on the clustering algorithm and the interpolation method; Obtaining the real-time coupling efficiency from the optimized optical path parameters.

7. The optical module production monitoring method according to any one of claims 1 to 6, characterized in that, The updating the thermal compensation model until the real-time coupling efficiency meets the preset requirements includes: Using the isolation forest algorithm to detect the abnormal data in the optical path parameters; Performing a deviation source analysis on the abnormal data, and optimizing the optical path parameters according to the deviation analysis result and in combination with the genetic algorithm; Iteratively updating the compensation coefficient of the thermal compensation model through the optimized optical path parameters until the updated real-time coupling efficiency meets the preset requirements.

8. An optical module production monitoring system based on artificial intelligence, characterized in that, Includes: An acquisition unit for acquiring the optical path parameters and temperature change data of the optical module to be detected; A calculation unit for calculating the fluctuation similarity between the time series corresponding to the optical path parameters and the temperature change data within the same time period, and determining the parameter drift trend of the optical path parameters according to the fluctuation similarity; A building unit for building a temperature field simulation model within a preset temperature range, and calculating the parameter drift distribution within the preset temperature range in combination with the parameter drift trend; A establishing unit for establishing a thermal compensation model based on the Gaussian process regression algorithm and the parameter drift distribution, and outputting the correction value of the optical path parameters; An adjustment unit for adjusting the optical path alignment according to the correction value and obtaining the real-time coupling efficiency; A judgment unit for judging whether the real-time coupling efficiency meets the preset requirements; An update unit for updating the thermal compensation model until the real-time coupling efficiency meets the preset requirements.

9. An electronic device, characterized in that, Includes: A central processing unit, a memory, an input / output interface, a wired or wireless network interface, and a power supply; The memory is a transient storage memory or a persistent storage memory; The central processing unit is configured to communicate with the memory and execute the instruction operations in the memory to execute the artificial intelligence-based optical module production monitoring method according to any one of claims 1 to 7.

10. A computer-readable storage medium includes instructions that, when run on a computer, cause the computer to execute the artificial intelligence-based optical module production monitoring method according to any one of claims 1 to 7.

Images