Data management system and method for AR glasses production quality tracing

Abstract

The invention relates to the technical field of intelligent manufacturing data management, in particular to a data management system and method for AR glasses production quality tracing. The method comprises the steps that in the active alignment curing process of an AR glasses module, position deviation, torque current and curing process data are synchronously collected at high frequency; based on the logarithmic coupling relationship between the position deviation and the torque current, calculating the tremor energy density representing the microscopic tremor; on the basis of the reciprocal weighting relation between the position deviation change rate and the curing remaining time, a curing drift damping index representing the later-period stability is calculated; and fusing the two indexes by using a normalized weight coefficient to obtain a comprehensive quality score, and carrying out dynamic quality grading and differential process processing on the module according to the comprehensive quality score. According to the method, the micro stress of the AR glasses module in the active alignment curing process is analyzed to evaluate the quality, and the accuracy of quality management tracing is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent manufacturing data management, in particular to a data management system and method for AR glasses production quality traceability. BACKGROUND

[0002] In the field of precision manufacturing of augmented reality (AR) and virtual reality (VR) devices, the assembly precision of the optical-mechanical module is a key factor that determines the optical performance of the final product and the user's visual experience. In order to achieve micron-level assembly precision, six-axis active alignment equipment is widely used in the industry for production. The working process of this type of equipment is usually to use a mechanical hand to hold the optical element, find the best optical position through image algorithm, and once the best coordinate is locked, the ultraviolet lamp is turned on to irradiate the photosensitive glue for curing to fix the position of the optical element.

[0003] The existing quality traceability management system mostly only records the final coordinate position and optical score at the moment of curing completion as result data when dealing with such precision processes. As long as these two values are within the range required by the specification, the system determines that the product is qualified.

[0004] However, this management method ignores the dynamic changes during the curing process and cannot capture the microscopic process of the motor continuously outputting torque to resist the volume shrinkage of the photosensitive glue under ultraviolet light irradiation. Although this microscopic change may not cause displacement to exceed the tolerance at the moment of curing completion, it will store a huge internal stress inside the module.

[0005] These products with high residual internal stress will slowly release internal stress after leaving the factory, causing a slight drift of the optical axis, which will eventually cause dizziness to the user during use. The existing technology lacks in-depth analysis of the curing process data and cannot distinguish between low-stress good products and high-stress hidden danger products, resulting in these products with potential defects flowing into the market, increasing the rate of after-sales repair.

[0006] Therefore, there is an urgent need for a method that can quantify the dynamic stability during the curing process, realize product life prediction and differentiated process management, and solve the problem of being unable to identify hidden high-stress products in the existing technology. SUMMARY

[0007] To solve the problem of the existing quality traceability management system ignoring the dynamic stress changes during the curing process and being unable to identify hidden products with high residual internal stress, the present application provides a data management system and method for AR glasses production quality traceability.

[0008] In a first aspect, the present application provides a data management method for AR glasses production quality traceability, comprising: During the active alignment and curing process of the AR glasses module, for each sampling moment, the position deviation, torque current and curing process data representing the remaining curing time of the AR glasses module are collected synchronously to obtain the position deviation sequence, torque current sequence and curing process sequence respectively. The flutter energy density is calculated by determining the coupling relationship between the position deviation sequence and the torque current sequence at the same sampling time. By analyzing the changes in position deviation at adjacent sampling times in the position deviation sequence as a function of curing process data in the curing process sequence, the curing drift damping index is determined. Based on the vibration energy density and the curing drift damping index, combined with the preset normalized weighting coefficient, the comprehensive quality score of the active alignment curing process is determined. The AR glasses module is dynamically graded according to the comprehensive quality score and automatically pushed to the corresponding differentiated processing route to realize data management for AR glasses production quality traceability.

[0009] This technical solution proposes a deep quality management system based on process data flow. Unlike existing technologies that only focus on the static results at the end of curing, this solution delves into the time dimension of the curing process. By calculating the vibration energy density, it can keenly capture the micro-oscillations when there is intense resistance between adhesive shrinkage and motor holding force, making the invisible internal stress accumulation process explicit. By calculating the curing drift damping index, combined with the time dimension, it penalizes the tiny displacements when the adhesive is close to solidification in the later stages of curing, accurately assessing the stability risks of the product throughout its entire life cycle. It enables refined graded management of products, intercepting or specially treating potentially high-stress products before they leave the factory. This not only effectively reduces the risk of optical axis drift and user dizziness caused by stress release, but also improves the accuracy of production quality traceability.

[0010] Preferably, for each sampling moment, the method for synchronously collecting the position deviation, torque current, and curing progress data representing the remaining curing time of the AR glasses module is as follows: The position deviation of the AR glasses module is collected by subtracting the encoder feedback register value from the set target register value from the motion control card; the position deviation represents the spatial micro-displacement of the AR glasses module relative to the target position at the current sampling moment; the torque current of the motor is read via the bus; the torque current represents the electromagnetic torque output by the motor to counteract adhesive shrinkage when maintaining the position of the AR glasses module; and the curing progress data of the AR glasses module is collected by calculating the time difference between the current sampling moment and the preset curing end moment; the curing progress data represents the remaining time of the active alignment curing process.

[0011] Preferably, the tremor energy density is determined based on the following relationship:

[0012] In the formula, The vibration energy density during the active alignment and curing process of the AR glasses module reflects the degree of high-frequency jitter during the active alignment and curing process. This represents the total number of sampling moments during the active alignment and curing process. This represents the sequence number of the sampling time. For the first The absolute value of the positional deviation at each sampling time. For the first Torque current at each sampling time, This is the preset base current correction constant. .

[0013] This technical solution constructs a physically meaningful energy density model, which amplifies the influence of micro-displacement by using the square term of the position deviation, and unifies the vibration in the positive and negative directions into energy potential. At the same time, it introduces logarithmically processed torque current as a weighting factor, reflecting the cost that the motor pays to maintain its position. When the motor outputs a large torque and the position still fluctuates, this indicator will rise sharply, thereby accurately identifying high-risk moments when the glue shrinks severely and internal stress remains seriously, providing an accurate basis for quality assessment.

[0014] Preferably, the curing drift damping index is determined based on the following relationship:

[0015] in, The curing drift damping index is used to reflect the positional instability during the active alignment and curing process of the AR glasses module. and The first The absolute value of the position deviation at the sampling time and the th sampling time The absolute value of the positional deviation at each sampling time. It is the absolute value symbol. For the first The remaining time of each sampling moment relative to the end time of the active alignment and curing process. The preset time decay smoothing factor is determined based on the curing rate characteristics of the photosensitive adhesive. This represents the total number of sampling times.

[0016] This technical solution is based on the physical law of the change of the curing state of glue with time in polymer chemistry. It constructs a time-reciprocal weighted cumulative model. As the curing process progresses, the remaining time gradually decreases and the denominator becomes smaller. This allows the model to give higher weight to the small displacements in the later stage of curing. This is because the displacement when the glue is close to solid can cause permanent lattice damage or microcracks. This index can effectively distinguish between the normal displacement in the fluid self-healing stage in the early stage of curing and the harmful displacement that causes structural damage in the late stage of curing, thereby accurately predicting the stability of the product.

[0017] Preferably, the basic current correction constant satisfies the following constraint: the sum of the basic current correction constant and the minimum value of the torque current is not less than 1, and the minimum value of the torque current is obtained by statistically analyzing historical production data of multiple AR glasses modules in advance.

[0018] Preferably, the overall quality score of the active alignment curing process is determined based on the following relationship: ;in, The overall quality score for the active alignment curing process. and All are preset normalized weighting coefficients. and These are the vibration energy density and curing drift damping index of the active alignment curing process of the AR glasses module, respectively.

[0019] This technical solution introduces weighting coefficients to uniformly map the vibration energy and drift index of different physical dimensions into score deduction items, thus constructing an intuitive quality score evaluation system. The more intense the vibration during the curing process or the more severe the drift in the later stage, the more points are deducted and the lower the score. This scoring mechanism transforms complex physical process data into quality levels that are easy to understand and implement, facilitating rapid decision-making on the production site and logical judgment of automated systems.

[0020] Preferably, the method for dynamically classifying the AR glasses module based on the comprehensive quality score is as follows: A first threshold, a second threshold, and a third threshold are preset in an increasing order, and first-class, second-class, third-class, and fourth-class products are preset in order of quality from best to worst; if the comprehensive quality score is greater than the first threshold, the AR glasses module is determined to be a first-class product; if the comprehensive quality score is not greater than the first threshold but greater than the second threshold, the AR glasses module is determined to be a second-class product; if the comprehensive quality score is not greater than the second threshold but greater than the third threshold, the AR glasses module is determined to be a third-class product; if the comprehensive quality score is not greater than the third threshold, the AR glasses module is determined to be a fourth-class product.

[0021] Preferably, the data is automatically pushed to the corresponding differentiated processing route to achieve data management for traceability of AR glasses production quality, including: for first-class products, the high-temperature aging process is waived and they go directly to the packaging line; for second-class products, the standard aging process is performed; for third-class products, they are marked as key observation objects and forced to undergo aging tests with double the duration; for fourth-class products, they are directly scrapped and not shipped.

[0022] This technical solution achieves risk classification management by implementing differentiated processing routes based on the comprehensive quality score of AR glasses modules. For ordinary products with slightly lower scores but still within the safety range, standard processes are implemented to ensure compliance. For high-risk products that have obvious vibrations but are not scrapped, aging tests of double the duration are enforced. This strategy can eliminate potential quality problems in the factory in a timely manner, preventing products with hidden defects from entering the market and causing customer complaints.

[0023] Preferably, after obtaining the position deviation sequence, torque current sequence, and curing process sequence, the method further includes: smoothing the position deviation sequence, torque current sequence, and curing process sequence using a sliding window mean filtering algorithm.

[0024] In a second aspect, the present invention also provides a data management system for traceability of AR glasses production quality, the system comprising a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the data management method as described in any one of the claims.

[0025] The present invention has the following effects: This invention makes the invisible internal stress in the AR glasses module production process explicit by evaluating the vibration energy density and curing drift damping index during the active alignment curing process, and performs differentiated processing based on dynamic grading. This effectively intercepts hidden defective products that cannot be identified by traditional detection methods and improves the accuracy of production quality traceability. Attached Figure Description

[0026] Figure 1 This is a flowchart of the present invention; Figure 2 This is a schematic diagram of real-time monitoring of the active alignment and solidification process of the AR glasses module of the present invention; Figure 3 This is a schematic diagram comparing the product quality screening capabilities of the present invention and existing technologies. Detailed Implementation

[0027] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0028] Referring to the technical solution of this invention, a data management method for quality traceability in AR glasses production specifically includes the following steps: S1: Obtain the position deviation sequence, torque current sequence, and curing process sequence of the AR glasses module during the active alignment and curing process.

[0029] In the assembly process of the optical engine module of AR glasses, in order to pursue micron-level optical performance, it is necessary to use active alignment equipment for precision assembly. The active alignment and curing process refers to the process of turning on the ultraviolet light source to irradiate the photosensitive adhesive to cause a cross-linking polymerization reaction while the robotic arm holds the optical element and adjusts it to the optimal optical position (active alignment stage) and keeps it stationary (UV curing stage).

[0030] When photosensitive adhesive changes from liquid to solid, it undergoes volume shrinkage, generating a huge internal pulling force that attempts to pull the optical element away from its optimal position. In order to keep the position coordinates of the optical element absolutely unchanged, the six-axis motor (a high-precision servo motor group that drives a six-degree-of-freedom adjustment platform) must continuously output reverse electromagnetic torque through a closed-loop control system to resist the mechanical force.

[0031] Existing quality management systems often only record the final coordinates after curing, ignoring the hidden details during the curing process. These micro-vibrations and stress accumulations during the process are the root cause of optical axis drift and dizziness in users after the product leaves the factory.

[0032] Therefore, this step aims to establish a dynamic data stream that captures the entire curing process, transforming the invisible mechanical resistance into visible digital signals, thus providing a precise physical basis for subsequent quantification of internal stress.

[0033] Specifically, the system first obtains the unique identification code of the current AR glasses module through a barcode scanner or RFID reader. This identification code will serve as the primary key for all subsequent process data and will be used to create an index in the database to ensure that each set of collected position deviation sequences and torque current sequences can be accurately associated with the specific physical product.

[0034] Set the sampling frequency to 1000Hz (i.e., 1000 samples per second), and simultaneously collect and lock the following three sets of time-series data throughout the entire time window from when the UV lamp is turned on to when curing is complete: The first set is the position deviation sequence, which reflects the microscopic displacement of the module in space. Specifically, it is obtained by subtracting the target theoretical position register value locked by the system from the encoder feedback register value of the motion control card at each sampling time. The difference is the position deviation, which can accurately reflect the degree to which the optical element deviates from the ideal position under the action of glue shrinkage force.

[0035] The second set is the torque-current sequence, which directly reflects the motor's effort to maintain its current position. Specifically, it is obtained by directly reading the motor's current value via the EtherCAT bus at each sampling moment. A higher current value indicates a greater reaction force on the motor, i.e., stronger adhesive contraction force. EtherCAT is an industrial-grade real-time Ethernet fieldbus protocol. Because EtherCAT has microsecond-level synchronous communication capabilities, it can bypass the multi-task scheduling delay of the host computer operating system and directly access the servo driver's underlying registers, thus ensuring strict alignment of current and position deviation data on the time axis.

[0036] The third group is the curing process sequence, which records the remaining time from the current sampling time to the end of curing. At each sampling time, the curing process data of the AR glasses module is collected by calculating the time difference between the current sampling time and the preset curing end time. The curing process data is used to characterize the remaining time of the active alignment curing process.

[0037] After acquiring the raw data, due to the complex electromagnetic environment in industrial sites, the raw signals are often mixed with high-frequency spike noise. In order to ensure the accuracy of subsequent analysis, this step uses a sliding window mean filtering algorithm to smooth the raw data. The window size is set to 5 sampling times. For the data of each dimension at each sampling time, the data and the data of the two sampling times before and after are taken, and the average value is calculated as the data of that sampling time. Through this smoothing operation, noise interference without physical meaning is eliminated, and a pure data stream that can truly reflect the physical process is obtained.

[0038] S2: Based on the position deviation sequence, torque current sequence and curing process sequence, determine the vibration energy density and curing drift damping index of the active alignment curing process of the AR glasses module.

[0039] After obtaining the clean process data stream, the raw data alone is not enough to directly judge the product quality, because a large positional deviation may be due to a huge external force or insufficient motor rigidity, while a large current may be due to a large adhesive shrinkage force or a heavy load itself.

[0040] To accurately assess the internal stress experienced by the module during the active alignment and curing process, this step takes a physics perspective and extracts two core features from high-frequency data: one is an energy index that reflects the intensity of micro-oscillations during the process, used to reflect the intensity of the conflict between motor torque and adhesive shrinkage force; the other is a trend index that reflects the instability of the position in the later stage of curing, used to assess the risk level in the shaping stage. These two indicators can comprehensively reveal the internal health status of the module.

[0041] Specifically, it includes: S21: The flutter energy density is determined by analyzing the coupling relationship between the position deviation sequence and the torque current sequence at the same sampling time.

[0042] Vibration energy density is used to assess whether there is high-frequency vibration at the microscopic level in the module during the curing process. The physical logic suggests that when the adhesive shrinks violently and the motor is not rigid enough to resist it, a dual oscillation of position and current will occur, which is the main source of residual stress.

[0043] In one embodiment, the tremor energy density is determined based on the following relationship:

[0044] In the formula, This represents the vibration energy density during the active alignment and curing process of the AR glasses module. It reflects the degree of high-frequency jitter during the active alignment and curing process; the higher the value, the more unstable the module is during the active alignment and curing process. This represents the total number of sampling moments during the active alignment and curing process. This represents the sequence number of the sampling time. For the first The absolute value of the positional deviation at each sampling time (unit: micrometers). The square is used here to convert the vibration deviation in the positive and negative directions into potential energy. Simultaneously, the squaring operation amplifies the impact of larger deviations on the results, making the vibration energy density more sensitive to large vibration deviations. For the first Torque current at each sampling time (unit: amperes). This is the preset base current correction constant. The basic current correction constant satisfies the following constraints: ,in, This is the preset base current correction constant. To obtain the minimum torque current value obtained from the historical production data of multiple AR glasses modules in advance, This involves performing a logarithmic transformation on the current value. It should be noted that all parameters in this equation are used as scalar values ​​during the calculation to eliminate dimensional differences and avoid calculation errors.

[0045] This formula multiplies the square of the position deviation by the logarithmically reduced current value, constructing a weighted energy model. If the torque current is large, it indicates the motor is struggling against the significant adhesive shrinkage force. In this case, if the position deviation is also small, it means the motor, while exerting effort, has maintained its position, and the oscillation energy is relatively controllable. However, if both the torque current and the position deviation are large, it indicates the motor is struggling to resist the adhesive shrinkage, and the system enters a state of violent oscillation. In this situation, the product term increases dramatically, leading to a significant increase in the final vibration energy density.

[0046] This formula, through the coupling calculation of position deviation and torque current, accurately captures the sampling moment of high stress and introduces... This is because the range of current variation is usually large, and logarithmic transformation can smooth out numerical differences. The settings satisfy This ensures that the logarithm is always greater than or equal to 1, avoiding mathematical errors and guaranteeing the physical meaning of the calculation.

[0047] To illustrate the calculation process of tremor energy density more clearly, a simple calculation example is given below: Assuming a sampling frequency of 1000Hz and a total active alignment and curing process of 5 seconds (curing time of 5 seconds), then the total number of sampling moments is... Select the first one Data at each sampling time: assuming the position deviation at that time... Micrometer, torque current Ampere, fundamental current correction constant .

[0048] The contribution of this sampling moment in the cumulative summation is calculated as follows:

[0049] To simplify the demonstration, let's assume that the contribution value of all other sampling times in the cumulative summation is also 4. If this indicator rises sharply, it means that the motor current is very high at a certain sampling moment, the force is large, and vibration has occurred, resulting in positional deviation.

[0050] S22: By analyzing the changes in the position deviation of adjacent sampling times in the position deviation sequence with the curing process data in the curing process sequence, the curing drift damping index is determined.

[0051] The curing drift damping index is used to measure the positional instability in the later stages of curing. According to the principles of polymer chemistry, in the early stages of curing, the glue is in a liquid state, and small positional deviations can be self-healed by the fluid. However, in the later stages of curing, the glue is close to a solid state, and any small positional deviation will cause permanent lattice damage or microcracks. Therefore, the later the curing time, the heavier the penalty for positional deviations should be.

[0052] In one embodiment, the curing drift damping index is determined based on the following relationship:

[0053] in, The curing drift damping index is used to reflect the degree of failure of the module's ability to resist positional drift during the curing stage. In other words, it reflects the positional instability of the module during the active alignment and curing process. The larger the value, the higher the risk of displacement in the later stage. and The first The position deviation at the sampling time and the first sampling time Positional deviation at each sampling time It is the absolute value symbol. Reflects the first The instantaneous drift change of the position deviation at each sampling time. For the first The remaining time (in seconds) of the sampling time relative to the end time of the active alignment and curing process, that is, the time remaining in the curing process sequence at the sampling time. The solidification process data collected at each sampling time, over time... It will gradually approach 0. The preset time decay smoothing factor is a constant greater than 0 to prevent the denominator from being 0. N is the total number of sampling times. This represents the sampling time number. It should be noted that all parameters in this formula are used in the calculation as scalar values ​​to eliminate dimensional differences and avoid calculation errors.

[0054] In this relation, This is determined based on the curing rate characteristics of photosensitive adhesives. Since the transition of the adhesive from a liquid state to a gel state and then to a glassy state is not instantaneous, there is a characteristic transition period. The introduction of ...

[0055] This relationship is a time-reciprocal weighted cumulative model, where the numerator is the change in positional deviation and the denominator is the remaining time. In the initial stage of curing... The denominator is relatively large, meaning the positional deviation occurring at this point contributes little to the curing drift damping exponent, consistent with the self-healing physical characteristic of liquid adhesives. As the curing process nears its end... As the denominator becomes very small, it decreases sharply, and the reciprocal increases sharply. At this point, if the same positional deviation occurs, its contribution to the solidification drift damping index will be greatly amplified.

[0056] By using time-reciprocal weighting, a non-linear penalty mechanism is implemented: the later the positional deviation occurs, the more severe the risk the system determines in terms of product quality. The introduction of this indicator not only prevents calculation errors with a denominator of zero, but also adjusts the growth curve of the time weight to better match the physical hardening process of the glue changing from liquid to solid. In this way, the indicator can keenly identify potential products that have undergone minor displacement during the critical curing stage (usually the later stage).

[0057] To more clearly illustrate the calculation process of the curing drift damping index, a simple calculation example is given below: Assuming in the first The sampling time and the first sampling time At each sampling time, the same positional deviation occurred, and the displacement... micrometers, setting ; If the first The sampling time is in the early stage of the active alignment and solidification process, and there are still [times] remaining until the end. Second, , No. The contribution of each sampling time point to the curing drift damping index of the active alignment curing process is: .

[0058] If the first The sampling time was at the end of the active alignment and solidification process, with [number] days remaining until the end. Second, , No. The contribution of each sampling time point to the curing drift damping index of the active alignment curing process is: .

[0059] It can be seen that for the same amount of displacement, the contribution to the curing drift damping index at the end of curing is 25 times that at the beginning of curing. This is consistent with the physical law that positional deviation at the end of curing will cause permanent lattice damage.

[0060] like Figure 2As shown, the real-time monitoring of the active alignment curing process is illustrated. The left vertical axis represents the positional deviation, and the right vertical axis represents the torque current. Observing the middle region of the figure (approximately 1.5 to 3.5 seconds), the process of the photosensitive adhesive undergoing a violent shrinkage reaction can be clearly seen. At this time, the positional deviation curve exhibits a high-frequency oscillation in a sinusoidal shape, indicating that the module has undergone micro-displacement under the pull of the adhesive. Simultaneously, in order to resist this pull and attempt to pull the module back to zero, the torque current curve rises sharply. This coupling phenomenon of large-scale positional oscillation and high current load output is the physical root cause of the sharp increase in vibration energy density. It reveals the internal stress accumulating inside the module, and simply monitoring the static coordinates of the curing end point cannot capture this dynamic counter-process. Observing the right end region of the figure, as the curing process nears its end, the remaining time approaches 0, and the adhesive is close to a glassy state. At this time, the positional deviation curve shows a slight unidirectional deviation trend. Although its absolute value (approximately 0.05 micrometers) is much smaller than the oscillation amplitude in the middle, since this is the critical period of curing and molding, according to the curing drift damping index calculation model of the present invention, this tiny displacement at the end will be given a very high weight penalty, because tiny displacement during the glue curing stage will lead to permanent structural damage or optical axis deflection risk.

[0061] S3: Based on the vibration energy density and the solidification drift damping index, combined with the preset normalized weighting coefficient, determine the comprehensive quality score of the active alignment solidification process.

[0062] After obtaining energy indicators that reflect the intensity of process oscillations and trend indicators that reflect later stability, these two indicators can transform invisible internal stress and potential risks into specific numerical indicators, thereby accurately capturing high stress moments and end-point drift risks.

[0063] This step further considers that the physical dimensions and numerical ranges of these two indicators are different. Directly using them for quality judgment is neither intuitive nor convenient for management. In order to achieve standardized quality control, this step aims to establish a fusion model that maps these two different physical characteristics into a unified percentage evaluation system. By introducing normalized weight coefficients, the complex physical data is transformed into a simple and clear quality score, so that each produced module has a quantifiable health label, providing a unique decision basis for subsequent automated grading processing.

[0064] In one embodiment, the overall quality score of the active alignment curing process is determined based on the following relationship:

[0065] in, The comprehensive quality score for the active alignment and curing process is a dimensionless numerical value ranging from 0 to 100, with a maximum score of 100. and These are all preset normalized weighting coefficients, whose function includes mapping physical quantities of different dimensions to dimensionless deduction values. and These represent the vibration energy density and curing drift damping index of the active alignment curing process of the AR glasses module, respectively. It should be noted that... and These are all numerical indicators, meaning they have undergone dimensionless processing, retaining only their numerical scalar values. It should be noted that all parameters in this formula are used in the calculation as their scalar values ​​to eliminate dimensional differences and avoid calculation errors.

[0066] In this relation, Used to measure tremor energy density The physical values ​​are converted into corresponding deduction values. Used to measure curing drift damping index The physical numerical values ​​are mapped to deduction values. These two weighting coefficients are derived through historical training. One specific method is to pre-acquire 100 modules, obtain the vibration energy density and solidification drift damping index of each module, and use principal component analysis or linear discriminant analysis to determine the contribution of the two indicators to the quality impact. The normalized contribution values ​​are then set as follows: and .

[0067] This formula uses a subtraction model, starting with a perfect score of 100 and deducting points based on existing defects. This represents the quality deduction caused by process oscillations; the more severe the oscillations, the lower the score. The larger the size, the more points are deducted. This represents the quality deduction caused by drift in later stages; the more severe the drift in later stages, the lower the score. The larger the size, the more points are deducted. This reflects the total quality loss, which is subtracted from 100 to obtain the final overall quality score.

[0068] During the active alignment and curing process of AR glasses modules, increases in either vibration energy density or curing drift damping index lead to a higher total deduction, resulting in a lower final overall quality score. This accurately reflects a decline in product quality. This scoring mechanism is not only intuitive but also dynamically adjustable. and It can flexibly adapt to the different sensitivity of different product models to oscillation or drift.

[0069] To more clearly illustrate the calculation process of the curing drift damping index, a simple calculation example is given below: Using the example data from step S2, assuming that after calculation, , Set weight coefficients , ; Quality deductions due to process vibration: Quality deductions due to drift in later stages: Final overall quality score: .

[0070] like Figure 3 As shown, this illustrates the product quality screening capabilities of the present invention and existing technologies. The background bar chart represents the judgment results of the existing technology. For the vast majority of samples, the existing technology indiscriminately classifies them as qualified products with a comprehensive quality score of 100. The foreground line graph represents the comprehensive quality score of the present invention. Taking sample number 9 as an example, under the existing technology's detection system, the final position coordinates of this sample at the moment of curing completion fell within the tolerance range, and therefore it was judged as a qualified product by the system. However, under the monitoring of the present invention, the comprehensive quality score of sample number 9 was only 58 points, and it was marked as a latent defect. This is because the present invention traced back its entire process data and found that although sample number 9 eventually returned to the origin, it experienced extremely violent oscillations and resistance in the middle of curing and had an unstable drift tendency at the end of curing. These dynamic instabilities that occurred during the active alignment curing process led to a significant decrease in the comprehensive quality score of the AR glasses module. The existing technology would treat this sample as a qualified product, affecting normal use, while the present invention intercepts the fault in advance, improving the accuracy of quality management.

[0071] Thus, by using a comprehensive quality scoring model, complex multidimensional physical characteristics are transformed into an intuitive single score, which can accurately reflect the true level of product quality, especially revealing hidden quality declines that cannot be detected by existing technologies.

[0072] S4: Dynamically classify based on comprehensive quality scores and automatically push to the corresponding differentiated processing routes to achieve data management for traceability of AR glasses production quality.

[0073] After obtaining a comprehensive quality score for each AR glasses module, the traditional approach often involves simply determining whether it is qualified or unqualified. However, this binary approach ignores the continuous distribution of product quality. A product with a score of 99 and a product with a score of 61 may both be qualified, but their internal stress states and potential lifespans are drastically different. If the same post-processing techniques are applied to all qualified products, such as using a uniform aging test duration, it can easily lead to a waste of time and resources on high-quality products and improper handling of potentially high-risk products.

[0074] Therefore, this step aims to implement a dynamic and refined hierarchical management strategy, which divides products into different levels based on their comprehensive quality scores and matches the most appropriate handling method to each level, thereby maximizing production efficiency while ensuring the quality of shipments.

[0075] Specifically, it includes: The first threshold is set to 90, the second threshold to 75, and the third threshold to 60 in an increasing order. The products are also set to first-class, second-class, third-class, and fourth-class in order of quality from best to worst, and are labeled as excellent, ordinary, high-risk, and scrap, respectively.

[0076] The system tags each produced AR glasses module based on a comprehensive quality score and automatically pushes the data to the MES system to achieve quality data management, specifically including: like The label "Excellent" indicates that the module is extremely stable during the curing process and has very low internal stress. This eliminates the need for a high-temperature aging process, allowing the module to flow directly into the packaging line, reducing unnecessary production cycles and lowering inventory costs.

[0077] like The label "normal" indicates that the module experiences slight vibrations, but is within the safe range. It undergoes a standard aging process (e.g., 12 hours) to ensure stable performance and compliance with standard shipping criteria.

[0078] like The label "high risk" indicates that the module's optical indicators are qualified, but the process is subject to severe vibrations. The aging test is doubled (e.g., 48 hours of thermal shock) to force the potential stress to be released in advance through a harsh environment. If the parameters drift after aging, they will be intercepted. If they remain stable, they will be downgraded and shipped or reserved for other uses.

[0079] like The label "scrap" indicates that the process was severely unstable and the item is scrapped. Even if it passes the current optical test, it will not be shipped to avoid it entering the market and causing potential risks.

[0080] Using the calculation results from S3, the overall quality score of the AR glasses module during the active alignment and curing process was 66 points, and it was labeled as a high-risk product. Therefore, the system will automatically route it to the double aging test process, instead of shipping it directly or only performing standard aging as in the traditional way.

[0081] In this way, dynamic hierarchical management not only enables the rapid release of high-quality products, improving production efficiency, but also strictly intercepts high-risk products, effectively reducing the after-sales return rate and achieving dual optimization of quality and efficiency.

[0082] Finally, after determining the comprehensive quality score and completing the grading, the system also performs a data archiving operation: the unique identification code of the AR glasses module, the calculated vibration energy density, the solidification drift damping index, the comprehensive quality score, the determined grade label, and the original position deviation sequence and torque current sequence are compressed and packaged to generate a data history of the solidification process of the AR glasses module. This data history is then uploaded to the factory's MES (Manufacturing Execution System) via industrial Ethernet and persistently stored for subsequent production quality traceability.

[0083] The present invention also provides a data management system for AR glasses production quality traceability. The system includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps of a data management method for AR glasses production quality traceability, so as to realize AR glasses production quality data management.

[0084] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A data management method for quality traceability in AR glasses production, characterized in that, include: During the active alignment and curing process of the AR glasses module, for each sampling moment, the position deviation, torque current and curing process data representing the remaining curing time of the AR glasses module are collected synchronously to obtain the position deviation sequence, torque current sequence and curing process sequence respectively. The flutter energy density is calculated by determining the coupling relationship between the position deviation sequence and the torque current sequence at the same sampling time. By analyzing the changes in position deviation at adjacent sampling times in the position deviation sequence as a function of curing process data in the curing process sequence, the curing drift damping index is determined. Based on the vibration energy density and the curing drift damping index, combined with the preset normalized weighting coefficient, the comprehensive quality score of the active alignment curing process is determined. The AR glasses module is dynamically graded according to the comprehensive quality score and automatically pushed to the corresponding differentiated processing route to realize data management for AR glasses production quality traceability.

2. The data management method according to claim 1, characterized in that, The method for synchronously collecting position deviation, torque current, and curing progress data representing the remaining curing time of the AR glasses module at each sampling moment is as follows: The position deviation of the AR glasses module is collected by subtracting the encoder feedback register value from the set target register value from the motion control card. This position deviation represents the microscopic spatial displacement of the AR glasses module relative to the target position at the current sampling moment. The torque current of the motor is read via the bus. This torque current represents the electromagnetic torque output by the motor to counteract adhesive shrinkage when maintaining the position of the AR glasses module. The curing progress data of the AR glasses module is collected by calculating the time difference between the current sampling moment and the preset curing end moment. This curing progress data represents the remaining time of the active alignment curing process.

3. The data management method according to claim 1, characterized in that, The tremor energy density is determined based on the following relationship: ； In the formula, The vibration energy density during the active alignment and curing process of the AR glasses module reflects the degree of high-frequency jitter during the active alignment and curing process. This represents the total number of sampling moments during the active alignment and curing process. This represents the sequence number of the sampling time. For the first The absolute value of the positional deviation at each sampling time. For the first Torque current at each sampling time, This is the preset base current correction constant. .

4. The data management method according to claim 1, characterized in that, The curing drift damping index is determined based on the following relationship: ,in, The curing drift damping index is used to reflect the positional instability during the active alignment and curing process of the AR glasses module. and The first The absolute value of the position deviation at the sampling time and the th sampling time The absolute value of the positional deviation at each sampling time. It is the absolute value symbol. For the first The remaining time of each sampling moment relative to the end time of the active alignment and curing process. The preset time decay smoothing factor is determined based on the curing rate characteristics of the photosensitive adhesive. This represents the total number of sampling times.

5. The data management method according to claim 3, characterized in that, The basic current correction constant satisfies the following constraint: the sum of the basic current correction constant and the minimum value of the torque current is not less than 1, and the minimum value of the torque current is obtained by statistically analyzing historical production data of multiple AR glasses modules in advance.

6. The data management method according to claim 1, characterized in that, The overall quality score of the active alignment curing process is determined based on the following relationship: ;in, The overall quality score for the active alignment curing process. and All are preset normalized weighting coefficients. The vibration energy density during the active alignment and curing process of the AR glasses module. The curing drift damping index is the curing drift damping index of the active alignment curing process of the AR glasses module.

7. The data management method according to claim 1, characterized in that, The method for dynamically classifying the AR glasses module based on the comprehensive quality score is as follows: The first, second, and third thresholds are preset in an increasing order, and the first, second, third, and fourth grades are preset in order of quality from best to worst. If the overall quality score is greater than the first threshold, the AR glasses module is determined to be a first-class product; if the overall quality score is not greater than the first threshold but greater than the second threshold, the AR glasses module is determined to be a second-class product; if the overall quality score is not greater than the second threshold but greater than the third threshold, the AR glasses module is determined to be a third-class product; if the overall quality score is not greater than the third threshold, the AR glasses module is determined to be a fourth-class product.

8. The data management method according to claim 7, characterized in that, Automatically push data to the corresponding differentiated processing route to achieve data management for AR glasses production quality traceability, including: For first-class products, the high-temperature aging process is waived, and they go directly to the packaging line; For second-grade products, the standard aging process shall be implemented. For third-grade products, they are marked as key observation targets, and aging tests of double the duration are enforced. Fourth-grade products will be scrapped immediately and will not be shipped.

9. The data management method according to claim 1, characterized in that, After obtaining the position deviation sequence, torque current sequence, and curing process sequence, the method further includes: smoothing the position deviation sequence, torque current sequence, and curing process sequence using a sliding window mean filtering algorithm.

10. A data management system for quality traceability in AR glasses production, characterized in that, The system includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the data management method as described in any one of claims 1-9.

Images