Production control method and device of electronic equipment, equipment and storage medium

Abstract

The invention belongs to the technical field of intelligent manufacturing, and discloses a production control method and device of electronic equipment, control equipment and a storage medium. The composite firmware package comprises production firmware and delivery firmware which are embedded in an independent partition mirror image form, so that hardware calibration and function test can be carried out on the electronic equipment in a running environment of the production firmware, and after the complete machine coupling of the electronic equipment is completed, a starting instruction is directly sent to the electronic equipment, so that the hardware calibration and function test can be carried out on the electronic equipment. Therefore, the electronic equipment automatically upgrades the production firmware to the delivery firmware. The starting instruction is instantly sent through the network and can be disconnected after being sent, online waiting for the upgrading process of the electronic equipment is not needed, simultaneous automatic upgrading of large-batch equipment can be achieved, the production efficiency is greatly improved, the composite firmware package is burnt at a time, re-burning is not needed, the burning process can be simplified, and the reject ratio is reduced.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of intelligent manufacturing, and particularly relates to a production control method and device of an electronic device, a control device and a storage medium. BACKGROUND

[0002] In the production and manufacturing process of electronic devices, especially original design manufacturer (ODM) devices, the production firmware needs to be built-in with software services strictly matched with the version of the production line test tool, while the delivery firmware is provided by the ODM customer, and the internal software service version may be incompatible with the production line tool, resulting in that the delivery firmware cannot be directly used for production test. Therefore, the production firmware containing specific test tools and drivers is usually used to complete the hardware calibration and function test of the device. After all production procedures are completed, the customer-specified delivery firmware is burned into the device through a burning tool to replace the production firmware.

[0003] At present, the process commonly used in the industry is as follows: first, the production firmware is batch burned into the device mainboard using a multi-machine download tool (usually through a USB interface) at the download station of the production line; second, the device mainboard completes all production test processes on the subsequent stations (such as calibration and single board test); and finally, a station is added at the end of the process, and the customer-specified delivery firmware is burned into the device again using the multi-machine download tool to overwrite the original production firmware.

[0004] The traditional technical solution adopts a twice-burning process, but the multi-machine download tool itself has a certain failure probability (for example, 1%) in burning. This means that each device will experience two failure risks, which theoretically doubles the whole machine failure rate caused by burning, increases the rework cost and production uncertainty. Moreover, the multi-machine download tool based on USB usually has a quantity limit (such as one-to-eight), and the operator must wait for the entire burning process to end before processing the next batch of devices, which becomes a bottleneck of production efficiency, resulting in a very low automation efficiency of the production line. SUMMARY

[0005] The purpose of the present application is to provide a production control method and device of an electronic device, a device and a storage medium, which can simplify the burning process, reduce the failure rate, and greatly improve the production efficiency.

[0006] The present application discloses a production control method of an electronic device, comprising: In the production process, a preset composite firmware package is burned into the mainboard of the electronic device through a burning tool; wherein the composite firmware package comprises a production firmware and a delivery firmware embedded in the form of an independent partition image; the production firmware is a firmware on the production line, and the delivery firmware is a firmware at the time of shipment; controlling the electronic device to start and run the production firmware, and performing hardware calibration and function test on the electronic device in the running environment of the production firmware; after the whole machine coupling of the electronic device is completed, sending a starting instruction to the electronic device to make the electronic device self-upgrade the production firmware to the delivery firmware.

[0007] In some embodiments, after sending the starting instruction to the electronic device to make the electronic device self-upgrade the production firmware to the delivery firmware, the method further comprises: judging whether the electronic device restarts normally and runs the delivery firmware; if the electronic device cannot restart normally or run the delivery firmware, determining that the electronic device is burned in poorly; collecting current environmental parameters on the production line, and calculating an environmental index value according to the current environmental parameters; determining a target adjustment model corresponding to the environmental index value from a plurality of adjustment models; determining a target feature vector according to current burning data on the production line, inputting the target feature vector into the target adjustment model, and obtaining a target control parameter on the production line; adjusting the current control parameter of the production line to the target control parameter.

[0008] In some embodiments, after determining that the electronic device is burned in poorly, the method further comprises: obtaining a plurality of sensing images collected by different image sensing devices on the production line for the electronic device; fusing a plurality of the sensing images to obtain a target image, and identifying the target image to obtain a product code of the electronic device; obtaining process parameters associated with the product code of the electronic device; extracting production feature data of each burner according to the process parameters, and predicting a target failure probability of each burner according to the production feature data; controlling the burners with the target failure probability reaching a specified probability to be automatically locked, and controlling subsequent mainboards on the production line to be distributed to the burners with the target failure probability lower than the specified probability.

[0009] In some embodiments, predicting a target failure probability of each burner according to the production feature data comprises: inputting the production feature data of each burner into a pre-trained prediction model to obtain a first failure probability of each burner; determining a target failure probability according to the first failure probability.

[0010] In some embodiments, determining a target failure probability according to the first failure probability comprises: According to the production feature data, the health index of each burner is predicted; According to the health index, a second failure probability is calculated; The average of the first failure probability and the second failure probability is determined as the target failure probability.

[0011] In some embodiments, the current environmental parameters at least include air temperature, average radiant temperature, air flow rate, relative humidity, air pollutant concentration, and equipment tolerable limit deviation; calculating an environmental index value according to the current environmental parameters comprises: According to the air temperature, the average radiant temperature, the air flow rate and the relative humidity, an air thermal comfort index is calculated; According to the air pollutant concentration, an air quality health index is calculated; According to the measured values of all current environmental parameters and the equipment tolerable limit deviation, a device operating environment comprehensive index is calculated; According to the air thermal comfort index, the air quality health index and the device operating environment comprehensive index, an environmental index value is calculated.

[0012] The second aspect of the present application discloses a production control device of an electronic device, comprising: A burning unit is used to burn a preset composite firmware package to the mainboard of the electronic device through a burning tool in the production process; wherein the composite firmware package comprises a production firmware and a delivery firmware embedded in the form of an independent partition image; the production firmware is the firmware on the production line, and the delivery firmware is the firmware when leaving the factory; A running unit is used to control the electronic device to start and run the production firmware, and to perform hardware calibration and function test on the electronic device in the running environment of the production firmware; An upgrading unit is used to send a start instruction to the electronic device after the whole machine coupling of the electronic device is completed, so that the electronic device upgrades the production firmware to the delivery firmware.

[0013] In some embodiments, the device further comprises: A judgment unit is used to judge whether the electronic device normally restarts and runs the delivery firmware after the upgrading unit sends a start instruction to the electronic device to make the electronic device upgrade the production firmware to the delivery firmware; A determination unit is used to determine that the electronic device is burned badly when the judgment unit determines that the electronic device cannot normally restart or run the delivery firmware; An acquisition unit is used to acquire the current environmental parameters on the production line; an environment computing unit configured to calculate an environment index value according to the current environment parameter; a calling unit configured to determine a target adjustment model corresponding to the environment index value from a plurality of adjustment models; a feature extraction unit configured to determine a target feature vector according to current burning data on the production line; a parameter optimization unit configured to input the target feature vector into the target adjustment model to obtain a target control parameter on the production line; a parameter adjustment unit configured to adjust a current control parameter of the production line to the target control parameter.

[0014] A third aspect of the present application discloses a control device, comprising a memory storing executable program codes and a processor coupled with the memory; the processor invokes the executable program codes stored in the memory to execute the production control method of the electronic device disclosed in the first aspect.

[0015] A fourth aspect of the present application discloses a computer readable storage medium storing a computer program, wherein the computer program causes a computer to execute the production control method of the electronic device disclosed in the first aspect.

[0016] The preset composite firmware package is burned to the mainboard of the electronic device by the burning tool in the production process, and the composite firmware package comprises a production firmware and a delivery firmware embedded in the form of an independent partition image, so that the hardware calibration and function test of the electronic device can be performed in the running environment of the production firmware, and after the whole machine coupling of the electronic device is completed, a start instruction is directly sent to the electronic device to make the electronic device upgrade the production firmware to the delivery firmware. Its beneficial effect lies in that the start instruction is sent instantaneously through the network, and the connection is disconnected after the instruction is sent, without online waiting for the upgrade process of the electronic device, so that the simultaneous automatic upgrade of a large number of devices can be realized, thereby greatly improving the production efficiency, and the composite firmware package is burned at one time, without the need for re-burning, so that the burning process can be simplified and the defective rate can be reduced. BRIEF DESCRIPTION OF DRAWINGS

[0017] The drawings herein show specific examples of the technical solutions of the present application, and constitute a part of the specification together with the specific embodiments, for explaining the technical solutions, principles and effects of the present application.

[0018] Unless specifically stated or defined otherwise, the same reference signs in different drawings represent the same or similar technical features, and different reference signs may also be used to represent the same or similar technical features.

[0019] Figure 1is a flow chart of a production control method of an electronic device disclosed by an embodiment of the present application; Figure 2 is a structural schematic diagram of a production control device of an electronic device disclosed by an embodiment of the present application; Figure 3 is a structural schematic diagram of a control device disclosed by an embodiment of the present application.

[0020] Explanation of reference signs: 201, a burning unit; 202, a running unit; 203, an upgrading unit; 301, a memory; 302, a processor. DETAILED DESCRIPTION

[0021] Unless specifically stated or otherwise understood, all technical and scientific terms used herein are understood to have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. In the context of the technical solutions of the present application in a real scenario, all technical and scientific terms used herein can also have meanings corresponding to the purposes of implementing the technical solutions of the present application. The terms "first, second, …" used herein are only used for distinguishing names and do not represent specific quantities or sequences. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0022] It should be noted that when an element is considered to be "fixed to" another element, it can be directly fixed to the other element or a middle element can exist; when an element is considered to be "connected to" another element, it can be directly connected to the other element or a middle element can exist; when an element is considered to be "mounted on" another element, it can be directly mounted on the other element or a middle element can exist. When an element is considered to be "provided on" another element, it can be directly provided on the other element or a middle element can exist.

[0023] Unless specifically stated or otherwise understood, "the", "this" used herein refers to the technical features or technical contents mentioned or described before the corresponding position, which can be the same as or similar to the technical features or technical contents mentioned. In addition, the terms "include" and "have" used herein and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units is not limited to the listed steps or units, but can optionally include steps or units not listed, or can optionally include other steps or units inherent to the process, method, product or device.

[0024] This invention discloses a production control method for electronic devices, which can be implemented through computer programming. The execution entity of this method is a production line control host, which can be configured as a control device such as a computer, laptop, or tablet, or a production control device embedded in the production line control host; this invention does not limit this. To facilitate understanding of this invention, specific embodiments are described in more detail below using a production line control host as an example, with reference to the accompanying drawings.

[0025] like Figure 1 As shown, the method includes the following steps 110-130: 110. During the production process, the production line control host uses a burning tool to burn a preset composite firmware package onto the motherboard of the electronic device; the composite firmware package includes production firmware and delivery firmware embedded in the form of independent partition images.

[0026] In this invention, the manufactured electronic device can be a smartwatch, smart earphone, or smartphone, etc. It should be noted that firmware refers to the device "driver" stored inside the device. Through firmware, the operating system can perform specific actions according to standard device drivers; for example, optical drives and CD burners all have internal firmware. Firmware is the software that performs the most basic and lowest-level functions of a system.

[0027] In this invention, "production firmware" is defined as firmware on the production line, used to drive electronic devices to complete all tests and calibrations in the production process; and "delivery firmware" is defined as firmware at the time of delivery, that is, the commercial version firmware finally delivered to the user.

[0028] In a specific embodiment of the present invention, a composite firmware package is pre-built, which physically contains the production firmware and embeds the final delivery firmware in the form of an independent partition image.

[0029] The specific implementation of pre-building the composite firmware package includes: when compiling the production firmware, checking whether the binary file of the delivered firmware (such as sysupgrade.bin) exists in the user-specified path through the compilation script (such as Makefile). If it exists, it is packaged into one or more independent partition images (such as user_data.img) and then integrated into the production firmware to generate the composite firmware package.

[0030] As an optional implementation, a partition strategy of the device flash memory can also be defined, the device flash memory is divided into at least three independent areas: a system partition (rootfs) for running a system, a preset partition (user_data) for storing a delivery firmware image file, and a calibration data partition (such as factory data) for storing calibration data. The calibration data partition is ensured to be independent of the system partition (rootfs) and the preset partition (user_data). Since the firmware self-upgrade process only covers the system partition, the integrity of the calibration data partition can be ensured, so that the calibration data can be inherited without loss.

[0031] In step 110, at the download station of the production line, a burning tool (such as a USB multi-machine download tool) can be used to burn the prepared composite firmware package to the mainboard of the electronic device at one time.

[0032] 120. The production line control host controls the electronic device to start and run the production firmware, and performs hardware calibration and function test on the electronic device in the running environment of the production firmware.

[0033] After the production line control host controls the electronic device to start, the production firmware system is run.

[0034] In the subsequent "calibration station" and "single board test station", the electronic device completes all hardware calibration and function test in the running environment of the production firmware. Among them, the key data generated in the process of hardware calibration and function test is written into the independent calibration data partition.

[0035] 130. After the whole machine coupling of the electronic device is completed, the production line control host sends a start instruction to the electronic device to make the electronic device self-upgrade the production firmware to the delivery firmware.

[0036] Specifically, in the whole machine test link after the whole machine coupling is assembled into a complete device, the production line control host remotely logs in the electronic device through a network (such as Ethernet) in a Telnet or SSH manner, and issues a start instruction (such as the sysupgrade command of OpenWrt) to the electronic device. After the electronic device receives the start instruction, a preset self-upgrade script (such as custom_FW_selfUpgrade.sh) is executed to complete the self-upgrade operation of upgrading the production firmware to the delivery firmware; wherein the core function of the self-upgrade script is to call the built-in upgrade instruction of the system and specify the upgrade source as the binary file of the delivery firmware preset in the delivery partition.

[0037] Further, the electronic device overwrites the content of the delivery firmware to the system partition (rootfs) in the process of performing the firmware self-upgrade. Since the "calibration data partition" is independent, the upgrade process does not affect any calibration data stored therein, achieving lossless inheritance of the calibration data. After the upgrade is completed, the electronic device automatically restarts, and the final "delivery firmware" system is run after the restart, and all calibration data generated in the production process is retained, becoming a qualified delivery product.

[0038] In summary, by implementing the embodiments of the present application, the traditional two-time physical burning is reduced to one time, and the production test and deployment of the final delivery firmware can be completed through only one physical burning, thereby fundamentally reducing the production failure rate caused by burning failure by 50%, significantly reducing the rework cost, and greatly reducing the production failure rate.

[0039] In addition, the second burning firmware upgrade operation is changed to an efficient triggering mechanism that can be triggered remotely, in parallel, and in batches through offline instructions (i.e., start instructions) over a network. Compared with the USB burning method which must be online and has a port number limit, the efficiency of the present method is higher, greatly releases the host computer resources, improves the automation level and flow speed of the production line, and thus greatly improves the production efficiency.

[0040] In addition, the production firmware and the delivery firmware are decoupled. The customer does not need to worry about whether the delivery firmware provided by the customer is compatible with the test tools on the production line, but only needs to provide the final product function. This reduces the development and maintenance burden of the customer, and also greatly reduces the firmware acceptance workload of the present application, without the need for complex compatibility traversal testing, thereby simplifying the firmware management and acceptance process.

[0041] Moreover, through reasonable partition design, it is ensured that the key production data (i.e., calibration data) generated on the production line during the firmware version switching process is completely retained, thereby ensuring the consistency and quality of the product.

[0042] Further, considering that the control in the production process of the electronic device in the prior art usually adopts fixed control parameters, lacking intelligent and self-adaptive adjustment capability, the production process is difficult to adapt to dynamic changes, affecting the production efficiency and product quality. Optionally, in other possible embodiments, after the step 130 is performed, the following non-illustrated steps 140-190 can also be performed: 140, determining whether the electronic device is normally restarted and runs the delivery firmware.

[0043] 150, if the electronic device cannot be normally restarted or run the delivery firmware, determining that the electronic device is burned badly.

[0044] Specifically, a check code (e.g. CRC, MD5 or SHA check code) can be embedded in the delivery firmware in advance. If the electronic device cannot be normally restarted or run the delivery firmware after being burned, the integrity of the embedded check code is automatically verified. When it is verified that the embedded check code is incomplete, it is determined that the electronic device is burned badly.

[0045] Optionally, after determining that the electronic device is burned badly, the number of times of burning badly of the burner for the electronic device is increased by one, and the number of times of burning badly of the same burner is monitored. If the number of times of burning badly of the same burner reaches a preset number within a specified time period (e.g. 15 minutes), the production line control host controls the burner to be automatically locked and stopped from burning; the production line control host can also send a notification message to a user terminal to remind an engineer to repair, and subsequent products are automatically routed to a backup burning station. In this way, by detecting and recording the number of times of burning badly, an abnormal alarm can be triggered, for example, the burning is automatically stopped if there are 3 times of burning badly within 15 minutes.

[0046] Further, after determining that the electronic device is burned badly, and before increasing the number of times of burning badly of the burner for the electronic device by one, it can be determined whether the electronic device is burned badly for the first time. If the electronic device is burned badly for the first time, the production line control host sends a retry instruction to the electronic device to make the electronic device automatically retry a specified number of times of self-upgrade operation; for example, the electronic device is automatically retried 2 times of self-upgrade operation if it is burned badly for the first time. After the electronic device automatically retries the specified number of times of self-upgrade operation, if it is determined again that the electronic device is burned badly, the number of times of burning badly of the burner for the electronic device is increased by one.

[0047] In this way, it can be determined that the burner is faulty only after the electronic device is burned badly for multiple times, thereby excluding the event of simply being burned badly for the first time, and improving the identification accuracy of the burner fault.

[0048] As an optional implementation, after determining that the electronic device is burned badly, the following steps 151-155 not shown in the figure can also be performed: 151. Obtain a plurality of sensing images collected by different image sensing devices on the production line for the electronic device.

[0049] A plurality of different sensing devices can be provided on the production line, such as a visible light camera, an infrared thermal imager, a multispectral scanner, a radar, etc.

[0050] 152. Fuse the plurality of sensing images to obtain a target image, and identify the target image to obtain a product code of the electronic device.

[0051] Fusing multiple sensor images can improve the information synthesis of multiple images captured for the same scene, and generate a "target image" (or fused image) with richer, more comprehensive and more reliable information. Different sensors have their own advantages and disadvantages, and the fused image can overcome the limitations of a single sensor. Through cross-verification of multiple source information, uncertainty is reduced and the confidence of interpreting the scene or target is improved. Specifically, fusing multiple sensor images to obtain a target image includes: Aligning multiple sensor images to the same coordinate system through affine transformation; performing multi-level discrete wavelet transform on each aligned image to obtain high frequency coefficients and low frequency coefficients of each sensor image; performing weighted fusion on the high frequency coefficients of multiple sensor images to obtain fused high frequency coefficients, and performing weighted fusion on the low frequency coefficients of multiple sensor images to obtain fused low frequency coefficients; and performing inverse discrete wavelet transform on the fused high frequency coefficients and low frequency coefficients to obtain a final fused image as a target image.

[0052] In the weighted fusion, different weight coefficients can be assigned to different sensor images, and the weight coefficients are assigned according to the local variance or information entropy of the image. The higher the local variance or information entropy of the image, the higher the weight coefficient assigned. Alternatively, the weight coefficients of different sensor images can also be determined according to the factory accuracy of different image sensors, for example, a sensor with a ±1% accuracy has a higher weight coefficient than a sensor with a ±5% accuracy. That is, the higher the accuracy, the higher the weight coefficient.

[0053] Optionally, after identifying the product code of the electronic device in step 152, the chip batch number can also be identified according to the product code of the electronic device; the defective rate of the chip batch number is calculated, and when the defective rate increases to a specified threshold within a specified time period, it is determined that the batch material is abnormal, and the production line control host sends an alarm information, which is used to notify the material department to check the batch material. In this way, by triggering an alarm when the defective rate increases rapidly in a short period of time, strict monitoring of the material can be realized.

[0054] 153、Obtain the process parameters associated with the product code of the electronic device.

[0055] The process parameters include, but are not limited to, burner number, station number, operator ID, burning time, burning duration, software version, firmware version, burning environment temperature, burning frequency of the burner, etc.

[0056] 154、According to the process parameters, extract the production characteristic data of each burner, and predict the target failure probability of each burner according to the production characteristic data.

[0057] Extract the production characteristic data of each burner, such as: recent failure rate, average burning time, frequency of occurrence of specific error codes, continuous production time, etc.

[0058] Optionally, in step 154, the specific implementation of predicting the target failure probability of each burner according to the production characteristic data can include: Input the production characteristic data of each burner into a pre-trained prediction model to obtain a first failure probability of each burner, and determine the target failure probability according to the first failure probability. The target failure probability refers to the probability of failure of the burner in the next 24 hours.

[0059] The prediction model can predict the probability of failure of the burner according to the input characteristics. The prediction model can be trained based on a machine learning model, for example, using a tree model such as XGBoost or LightGBM for training.

[0060] In training the prediction model, historical characteristic data X of each burner at each time point (including normal and pre-failure data) is obtained, such as [failure rate, average time, error code A times, error code B times, running cycle number, …], and the corresponding label y is configured. The label y is used to define the time point of "failure". For example, the data within N hours (such as 24 hours) before the failure is marked as 1 (failure), and the stable running data much earlier than the failure is marked as 0 (no failure). The data of the device without failure is always marked as 0. Then the historical characteristic data X is input into the tree model, so that the model learns to output the probability of being labeled as 1 (failure), and a binary classification cross-entropy loss function (Log Loss) is used to iteratively optimize the model.

[0061] Further preferably, the health index of each burner can be predicted according to the production characteristic data, and a second failure probability can be calculated according to the health index. The average of the first failure probability and the second failure probability is determined as the target failure probability. Compared with directly taking the first failure probability as the target failure probability, the correction according to the health index can further improve the accuracy of failure prediction.

[0062] The implementation of predicting the health index of each burner according to the production characteristic data can include: The production characteristic data of each burner is input into a pre-trained anomaly detection model to obtain an anomaly score of each burner; and a health index of each burner is calculated according to the anomaly score of each burner. The anomaly detection model is a model trained only using data of all burners during a normal stable operation period, such as an Isolation Forest or a One-Class SVM machine learning model. By learning the data pattern of the device in the normal state, the running state deviating from the pattern is regarded as "abnormal", and the greater the deviation, the lower the health index.

[0063] Specifically, the health index of each burner is calculated as follows: After normalizing the anomaly score, subtract 100 to obtain the health index HI: HI = max(0, 100 - (Anomaly_Score - Score_min) / (Score_max - Score_min) * 100); where Anomaly Score is the anomaly score, Score_min is the minimum value of the anomaly score of all burners, and Score_max is the maximum value of the anomaly score of all burners.

[0064] It should be noted that the second failure probability is negatively correlated with the health index, and the higher the health index, the lower the corresponding second failure probability, and the sum of the two is 1. For example, a health index of 0.3 corresponds to a second failure probability of 70%.

[0065] 155、The production line control host controls the burners whose target failure probability reaches a specified probability to be automatically locked, and controls the subsequent mainboards on the production line to be distributed to burners whose target failure probability is lower than the specified probability.

[0066] In this way, the subsequent mainboards to be burned can be distributed to burners with high health indexes, and devices with low health indexes can be controlled to suspend or reduce production, and maintenance work orders can be automatically generated. By no longer regarding each burning failure event as an independent accidental event, but associating it with specific device codes and production process data, and through calculation and analysis, the change from "after-the-fact remediation" to "before-the-fact prediction and during-the-process intervention" is achieved, so that a data-driven intelligent optimization system can be realized, and the intelligent degree of electronic device production control can be improved.

[0067] 160、Collect the current environmental parameters on the production line, and calculate an environmental index value according to the current environmental parameters.

[0068] The current environment parameters on the production line include, but are not limited to, environmental data such as temperature, humidity, light intensity, air quality index, oxygen content, carbon dioxide content, particle size distribution, wind speed, etc. in the workshop, and various machine running parameters, sensor collection frequency or accuracy, etc. on the production line.

[0069] Specifically, in step 160, the current environment parameters at least include air temperature, mean radiant temperature, air flow rate, relative humidity, air pollutant concentration, and equipment tolerable limit deviation; then the environment index value is calculated according to the current environment parameters, which can include the following steps 161~164 not shown in the figure: 161、According to the air temperature, mean radiant temperature, air flow rate and relative humidity in the current environment parameters, the air thermal comfort index is calculated.

[0070] Among them, the core input variables when calculating the thermal comfort index (Predicted Mean Vote, PMV) are four environment parameters: air temperature (°C); mean radiant temperature (°C), usually approximated to air temperature; air flow rate (m / s); relative humidity (%). The PMV value is between -3 and +3, representing 7 feelings from "cold" to "hot". PMV=0 indicates thermal neutrality, the most comfortable.

[0071] 162、According to the air pollutant concentration in the current environment parameters, the air quality health index is calculated.

[0072] First, for each pollutant in the air (such as CO2, PM 2.5 , total volatile organic compounds (Total Volatile Organic Compounds, TVOC)), a separate index IAQI sub is calculated: IAQI sub = (IAQI high - IAQI low ) / (C high - C low ) * (C - C low ) + IAQI low ; Where C is the measured value of the pollutant concentration; C low , C high are the lower and upper limits of the pollutant concentration interval; IAQI low , IAQI high correspond to the lower and upper limits of the index value of C low and C high .

[0073] Then, the IAQI subValue: The maximum value among all pollutant sub-indices in the air is taken as the current Indoor Air Quality Index (IAQI).

[0074] 163. Calculate the comprehensive index of the equipment's operating environment based on the measured values ​​of all current environmental parameters and the limit deviations that the equipment can withstand.

[0075] First, obtain the ideal range for each measured value in the current environmental parameters, for example, the temperature requirement T. ideal ±ΔT, humidity requirement RH ideal ±ΔRH. Next, the deviation of each measurement value is calculated based on the ideal range, and the deviation of each measurement value is mapped to the interval [0, 1], where 0 represents the ideal and 1 represents the deviation from the tolerance range.

[0076] Deviation T = max(0, min(1, (|T) current - T ideal | -ΔT) / (T tolerance -ΔT) ));where, T current T is the measured value. ideal The ideal value is T, where ΔT is the deviation threshold. tolerance This refers to the maximum tolerance that the equipment can withstand.

[0077] Finally, the deviations of all measured values ​​are weighted and fused to calculate a comprehensive equipment operating environment index. This index characterizes the "gap" between the current environmental parameters and the ideal environmental parameters of the equipment. The closer the comprehensive operating environment index is to 0, the more suitable the environment is for equipment operation; the closer it is to 1, the higher the warning risk, and shutdown protection may be required. The weight coefficient of each measured value is determined according to its importance to the equipment, and the sum of the weight coefficients of all measured values ​​is one.

[0078] 164. Calculate the environmental index values ​​based on the air thermal comfort index, air quality health index, and comprehensive equipment operating environment index.

[0079] Specifically, a fuzzy set (such as "comfortable" and "uncomfortable"; "excellent air quality" and "poor air quality") can be defined for each sub-indicator. Then, the value of each sub-indicator can be converted into fuzzy concepts such as "good", "moderate", and "poor". The reasoning can then be performed using a preset rule base, and the reasoning result can be converted back into a precise value, which is the environmental indicator value.

[0080] 170. Determine the target regulation model corresponding to the environmental indicator value from multiple regulation models.

[0081] Considering that the requirements for the production line control parameters are different under different production environments, for example, in the case of too much humidity in the workshop, the data transmission rate during burning needs to be appropriately accelerated to prevent the equipment components from being affected by humidity and thus affecting the final yield. Therefore, in the embodiments of the present application, different adjustment models are configured for different environmental index values, which are equivalent to different optimization systems. Thus, different adjustment models can be adaptively called for control parameter optimization under different environments.

[0082] Specifically, the current environment level can be determined according to the environmental index value, and then one adjustment model corresponding to the current environment level is selected from a plurality of pre-trained adjustment models as a target adjustment model. For example, when the environmental index value is 15, it is determined to be in the range [0, 30), and the corresponding environment level is level one; when the environmental index value is 40, it is determined to be in the range [30, 60), and the corresponding environment level is level two; and so on. At each environment level, one or more adjustment models can be stored correspondingly. When there are multiple corresponding adjustment models, one of them is randomly selected, or an adjustment model that meets the user's needs is selected. Optionally, when there are multiple corresponding adjustment models, the production line control host can pop up multiple options on the user interface, and the multiple options correspond one-to-one to the multiple adjustment models, and then receive a selection instruction input by the user to determine the target adjustment model selected by the user according to the selection instruction.

[0083] Among them, different adjustment models can be pre-trained based on different machine learning models, such as deep learning, linear regression, support vector machine, decision tree, neural network, etc. Specifically, the training of the adjustment model of the production line control parameter can include the following steps S1-S3 which are not shown in the figure: S1, obtaining historical burning data on the electronic device production line, and obtaining sample feature vectors at multiple sampling times by feature extraction based on the historical burning data.

[0084] Among them, the historical burning data can be the burning data in a historical period (such as the past week, the past half month), and the sample feature vector is the same as the target feature vector, including a vector composed of multiple feature values such as burning parameters, cumulative failure rate, and verification failure rate. Interval sampling means determining a plurality of discrete sampling times from the historical period at a preset time interval, and obtaining sample feature vectors corresponding to each sampling time.

[0085] Obtaining sample feature vectors at multiple sampling times by feature extraction based on historical burning data can include: variance of each feature in the current burning data is calculated; a feature column with variance greater than a preset threshold is reserved as a candidate feature; the importance of each candidate feature is calculated and sorted from large to small, and the top designated number of features in the importance from large to small are selected as principal components; feature extraction is performed on the historical burning data according to the principal components to obtain sample feature vectors at multiple sampling times.

[0086] wherein, for a single feature X j = [x1, x2, …, x n ], n is the number of samples, and the variance is calculated as: ; wherein, is the value of the i-th sample on the feature, is the mean value of all samples of the feature. The greater the variance, the more obvious the fluctuation of the feature value among different samples, and the more useful information it may contain; the variance close to zero indicates that the feature is almost a constant.

[0087] wherein, the principal component represents the main and key feature, and by screening out the representative feature, the dimension of the data can be reduced, and the reliability, accuracy and interpretability of the prediction result can be improved.

[0088] S2, obtain the corresponding optimization control parameter at each sampling time as label data.

[0089] The burning parameters on the production line directly affect the yield of the electronic equipment produced, and at each sampling time, the ideal optimization control parameter can be calculated as the label data for model training.

[0090] S3, according to the sample feature vector at each sampling time and the corresponding label data, the constructed machine learning model is trained to obtain the adjustment model of the production line control parameter.

[0091] Specifically, during training, the sample feature vector is divided into a training set and a validation set. First, the sample feature vector in the training set is input into the constructed machine learning model, and the predicted control parameter corresponding to each sample feature vector output by the machine learning model is obtained. Then, the loss value between the predicted control parameter and the label data is calculated, and the weight parameter of the machine learning model is iterated according to the loss value, until the performance index of the machine learning model on the validation set reaches a specified performance threshold, and the training is completed to obtain the adjustment model of the production line control parameter.

[0092] 180, determine the target feature vector according to the current burning data on the production line, input the target feature vector into the target adjustment model, and obtain the target control parameter on the production line.

[0093] 190, adjust the current control parameter of the production line to the target control parameter.

[0094] wherein the current burning data refers to the burning data at the current time on the production line, and the target feature vector includes feature values of a plurality of specified features, such as values of specified features of the burning parameters, the cumulative failure rate, the verification failure rate, etc., which can be determined according to the current burning data. The cumulative failure rate and the verification failure rate can be calculated according to the number of defective products of the electronic device burning failure; and the burning parameters include but are not limited to the clock frequency, the timing (such as the SWD clock delay of STMicro), the burning voltage (the power supply voltage of the burning interface (such as UART, SWD, JTAG), the burning voltage fluctuation, the threshold value of high and low level (such as the matching of TTL, CMOS level), the data transmission rate (such as the baud rate), the storage path of the composite firmware package, the burning address, the erasing mode (full chip erasing, sector erasing or retaining part of the data (such as Bootloader)), the burning mode (full automatic, differential burning (only writing the changed part) or forced burning) and the like.

[0095] The target control parameter includes not only the burning parameter optimized by the adjusted model, but also various machine running parameters, sensor collection frequency or accuracy and the like on the production line.

[0096] Through the above steps 140-190, when the electronic device burning failure is determined in the production process, the target feature vector can be calculated according to the current burning parameter, and the optimized burning parameter can be predicted through the machine learning model, so as to realize the real-time dynamic adjustment of the burning parameter. Compared with the fixed control parameter, the intelligent control capability of the production line can be improved, and dynamic adaptation to the production process can be realized.

[0097] As shown in Figure 2 The embodiment of the present application discloses a production control device of an electronic device, which comprises a burning unit 201, a running unit 202 and an upgrading unit 203, wherein, The burning unit 201 is used for burning a preset composite firmware package to the mainboard of the electronic device through a burning tool in the production process; wherein the composite firmware package comprises a production firmware and a delivery firmware embedded in the form of an independent partition image; the production firmware is a firmware on the production line, and the delivery firmware is a firmware at the time of delivery; The running unit 202 is used for controlling the electronic device to start and run the production firmware, and performing hardware calibration and function test on the electronic device in the running environment of the production firmware; The upgrading unit 203 is used for sending a starting instruction to the electronic device after the whole machine coupling of the electronic device is completed, so as to make the electronic device upgrade the production firmware to the delivery firmware.

[0098] As an optional implementation, the production control device further comprises the following units which are not shown in the figure: The judgment unit is configured to judge whether the electronic device is restarted normally and runs the delivery firmware after the upgrading unit sends the starting instruction to the electronic device to make the electronic device self-upgrade the production firmware to the delivery firmware. The determination unit is configured to determine that the electronic device is burned in abnormally when the judgment unit determines that the electronic device cannot be restarted normally or run the delivery firmware. The acquisition unit is configured to acquire a current environmental parameter on the production line. The environmental calculation unit is configured to calculate an environmental index value according to the current environmental parameter. The calling unit is configured to determine a target adjustment model corresponding to the environmental index value from a plurality of adjustment models. The feature extraction unit is configured to determine a target feature vector according to current burning data on the production line. The parameter optimization unit is configured to input the target feature vector into the target adjustment model to obtain a target control parameter on the production line. The parameter adjustment unit is configured to adjust a current control parameter of the production line to the target control parameter.

[0099] As an optional implementation, the production control device further includes the following units not shown in the figure: The acquisition unit is configured to acquire a plurality of sensing images collected by different image sensing devices on the production line for the electronic device after the determination unit determines that the electronic device is burned in abnormally. The fusion unit is configured to fuse the plurality of sensing images to obtain a target image. The recognition unit is configured to recognize the target image to obtain a product code of the electronic device. The association unit is configured to acquire a process parameter associated with the product code of the electronic device. The extraction unit is configured to extract production feature data of each burner according to the process parameter. The prediction unit is configured to predict a target failure probability of each burner according to the production feature data. The control unit is configured to control the burners with the target failure probability reaching a specified probability to be automatically locked, and control subsequent motherboards on the production line to be distributed to the burners with the target failure probability lower than the specified probability.

[0100] As an optional implementation, the prediction unit described above includes the following sub-units not shown in the figure: The prediction sub-unit is configured to input the production feature data of each burner into a pre-trained prediction model to obtain a first failure probability of each burner. The determination sub-unit is configured to determine the target failure probability according to the first failure probability.

[0101] Further optionally, the determination sub-unit described above includes the following modules not shown in the figure: a prediction module configured to predict a health index of each burner according to the production feature data; a calculation module configured to calculate a second failure probability according to the health index; a correction module configured to determine an average value of the first failure probability and the second failure probability as a target failure probability.

[0102] As shown in Figure 3 The embodiment of the present application discloses a control device, comprising a memory 301 storing executable program codes and a processor 302 coupled with the memory 301. The processor 302 calls the executable program codes stored in the memory 301 to execute the production control method of the electronic device described in each of the above embodiments.

[0103] The embodiment of the present application also discloses a computer readable storage medium storing a computer program, wherein the computer program enables the computer to execute the production control method of the electronic device described in each of the above embodiments.

[0104] The above embodiments are intended to exemplarily reproduce and deduce the technical solutions of the present application, and to completely describe the technical solutions, purposes and effects of the present application, so as to make the public more thoroughly and comprehensively understand the disclosed content of the present application, and not to limit the protection scope of the present application.

[0105] The above embodiments are not based on the exhaustive enumeration of the present application, and there can be many other unlisted embodiments. Any replacement and improvement made without violating the concept of the present application shall fall within the protection scope of the present application.

Claims

1. A production control method of an electronic device, characterized by, The method comprises the following steps: In the production process, a preset composite firmware package is burned into the mainboard of the electronic device by a burning tool; wherein, the composite firmware package comprises a production firmware and a delivery firmware embedded in the form of an independent partition image; the production firmware is a firmware on the production line, and the delivery firmware is a firmware at the time of delivery; controlling the electronic device to start and run the production firmware, and performing hardware calibration and function test on the electronic device in the running environment of the production firmware; after the completion of the whole machine coupling of the electronic device, sending a start instruction to the electronic device to make the electronic device upgrade the production firmware to the delivery firmware.

2. The production control method of an electronic device according to claim 1, characterized by, After sending the start instruction to the electronic device to make the electronic device upgrade the production firmware to the delivery firmware, the method further comprises: determining whether the electronic device restarts normally and runs the delivery firmware; if the electronic device cannot restart normally or run the delivery firmware, determining that the electronic device is burned badly; collecting current environmental parameters on the production line, and calculating an environmental index value according to the current environmental parameters; determining a target adjustment model corresponding to the environmental index value from a plurality of adjustment models; determining a target feature vector according to the current burning data on the production line, inputting the target feature vector into the target adjustment model, and obtaining a target control parameter on the production line; adjusting the current control parameter of the production line to the target control parameter.

3. The production control method of an electronic device according to claim 2, wherein After determining that the electronic device is burned badly, the method further comprises: obtaining a plurality of sensing images collected by different image sensing devices on the production line for the electronic device; fusing a plurality of sensing images to obtain a target image, and identifying the target image to obtain a product code of the electronic device; obtaining process parameters associated with the product code of the electronic device; extracting production feature data of each burner according to the process parameters, and predicting a target failure probability of each burner according to the production feature data; controlling the burners whose target failure probability reaches a specified probability to be automatically locked, and controlling subsequent mainboards on the production line to be distributed to the burners whose target failure probability is lower than the specified probability.

4. The production control method of an electronic device according to claim 3, characterized by, The method for predicting the target failure probability of each burner according to the production feature data comprises: inputting the production feature data of each burner into a pre-trained prediction model to obtain a first failure probability of each burner; determining a target failure probability according to the first failure probability.

5. The production control method of an electronic device according to claim 4, wherein The method for determining a target failure probability according to the first failure probability comprises: predicting a health index of each burner according to the production feature data; calculating a second failure probability according to the health index; determining the average value of the first failure probability and the second failure probability as the target failure probability.

6. The production control method of an electronic device according to any one of claims 2 to 4, characterized by The current environmental parameters at least include air temperature, average radiation temperature, air flow rate, relative humidity, pollutant concentration in air, and limit deviation that the device can withstand; The method for calculating an environmental index value according to the current environmental parameters comprises: calculating an air thermal comfort index according to the air temperature, the average radiation temperature, the air flow rate, and the relative humidity; According to the concentration of pollutants in the air, an air quality health index is calculated; According to the measurement values of all current environmental parameters and the limit deviation that the device can withstand, a device operating environment comprehensive index is calculated; According to the air thermal comfort index, the air quality health index and the device operating environment comprehensive index, an environmental index value is calculated.

7. An apparatus for production control of an electronic device, characterized by comprising: Comprise: The burning unit is used to burn the preset composite firmware package to the mainboard of the electronic device through the burning tool in the production process; wherein, the composite firmware package comprises production firmware and delivery firmware embedded in the form of independent partition image; the production firmware is the firmware on the production line, and the delivery firmware is the firmware when leaving the factory; The running unit is used to control the electronic device to start and run the production firmware, and to perform hardware calibration and function test on the electronic device in the running environment of the production firmware; The upgrading unit is used to send a start instruction to the electronic device after the whole machine coupling of the electronic device is completed, so that the electronic device upgrades the production firmware to the delivery firmware.

8. The production control apparatus of an electronic device according to Claim 7, wherein Also include: The judgment unit is used to judge whether the electronic device normally restarts and runs the delivery firmware after the upgrading unit sends a start instruction to the electronic device to make the electronic device upgrade the production firmware to the delivery firmware; The determination unit is used to determine that the electronic device is not good when the judgment unit determines that the electronic device cannot normally restart or run the delivery firmware; The acquisition unit is used to acquire the current environmental parameters on the production line; The environmental calculation unit is used to calculate the environmental index value according to the current environmental parameters; The calling unit is used to determine the target adjustment model corresponding to the environmental index value from a plurality of adjustment models; The feature extraction unit is used to determine the target feature vector according to the current burning data on the production line; The parameter optimization unit is used to input the target feature vector into the target adjustment model to obtain the target control parameter on the production line; The parameter adjustment unit is used to adjust the current control parameter of the production line to the target control parameter.

9. A control device, characterized by The memory stores executable program code, and the processor is coupled with the memory; the processor invokes the executable program code stored in the memory, and is used to execute the production control method of the electronic device in any one of claims 1 to 6.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, wherein the computer program makes the computer execute the production control method of the electronic device in any one of claims 1 to 6.

Images