A chip online upgrade method, chip device and computer readable storage medium

Abstract

The present invention proposes a chip online upgrade method, a chip device and a computer-readable storage medium, wherein the method comprises: the chip establishes a network connection with an upgrade server, the chip sends the current firmware version information to the upgrade server, and checks whether a new update package is available; if a new update package is available, the upgrade server segments the update package and sends it to the chip via the network, and the upgrade server adjusts the size of the segments in real time according to the network performance; the upgrade package is transmitted based on the adjusted segment size, a segment transmission log is obtained, and whether the chip will have an upgrade failure is predicted based on the segment transmission log. Through this method, the transmission efficiency and reliability during the chip upgrade process can be improved.

Description

Technical Field

[0001] The present invention proposes a chip online upgrade method, a chip device and a computer-readable storage medium, and relates to the technical field of chip upgrade. Background Art

[0002] Over-the-Air (OTA) chip upgrade technology has become a key feature of modern electronic devices, especially in the fields of Internet of Things (IoT), smart home appliances, mobile devices and industrial automation. Through online upgrades, device manufacturers can remotely distribute firmware updates, patch security vulnerabilities, add new features or optimize existing performance. Compared with traditional firmware update methods, OTA technology has many advantages, including but not limited to improving user experience, reducing maintenance costs, and improving security and reliability during the product life cycle.

[0003] Although OTA upgrades have brought great convenience, they still face many challenges in their application, including: Fluctuations in network conditions: In traditional upgrade methods, the download speed and success rate of the fixed-segment size download strategy are seriously affected when the network status is unstable or the bandwidth is limited; many embedded devices and IoT devices have limited resources (such as memory and CPU), and are particularly prone to performance bottlenecks when processing large firmware packages; failures that occur during the upgrade process cannot be predicted in advance, thus affecting users' use of the device. Summary of the invention

[0004] The present invention provides a chip online upgrade method, a chip device and a computer-readable storage medium to solve the above-mentioned problems:

[0005] The present invention provides a chip online upgrade method, the method comprising:

[0006] The chip establishes a network connection with the upgrade server, and the chip sends the current firmware version information to the upgrade server to check whether there is a new update package available;

[0007] If a new update package is available, the update server will fragment the update package and send it to the chip via the network. The update server will adjust the fragment size in real time according to the network performance;

[0008] Transmitting the upgrade package based on the adjusted slice size, obtaining the slice transmission log, and predicting whether the chip will have an upgrade failure based on the slice transmission log;

[0009] Furthermore, the chip establishes a network connection with the upgrade server, and the chip sends the current firmware version information to the upgrade server to check whether a new update package is available, including:

[0010] The chip prepares current firmware version information, wherein the current firmware version information includes a version number and a release time;

[0011] The chip sends an HTTP request to the upgrade server to obtain the latest firmware version information;

[0012] The chip compares the current version with the latest version information returned by the upgrade server to determine whether an update is required.

[0013] Furthermore, the chip establishes a network connection with the upgrade server, and the network connection includes Wi-Fi, mobile network, Ethernet, LPWAN, Bluetooth and Zigbee.

[0014] Furthermore, if a new update package is available, the update server segments the update package and sends it to the chip via the network. The update server adjusts the size of the segments in real time according to the network performance, including:

[0015] If a new update package is available, the upgrade server compares the size of the update package to see if it exceeds a preset threshold, and if the size of the update package exceeds the preset threshold, the update package is fragmented;

[0016] The upgrade server adjusts the slice size according to the network performance through an adaptive slice model.

[0017] Furthermore, the adaptive sharding model is:

[0018]

[0019] in, represents the fragment size at the i-th transmission time, represents the instantaneous bit error rate at the i-th moment, represents the instantaneous network jitter at the i-th moment, represents the historical transmission failure rate at the i-th moment, represents the instantaneous equipment load factor at the i-th moment, represents the correction factor of the device resource limit at the i-th moment, B represents the base shard size, represents the maximum tolerable bit error rate, Indicates the maximum tolerable network jitter, The length of the time window representing the historical transmission failure rate, Indicates the network bandwidth. Indicates the start time of the time window. Indicates the end time of the time window.

[0020] Furthermore, the time window length of the historical transmission failure rate is calculated by a time window model. Specifically, the time window model is:

[0021]

[0022] in, represents the weighted difference between the current and past historical transmission failure rates, CV(H) represents the coefficient of variation of the historical transmission failure rate, represents the exponentially decaying weighted average of the historical transmission failure rate, Indicates the preset minimum time window length, Indicates the preset maximum time window length, represents the basic time window length correction factor, , and Represents the power exponent for adjusting the weights of different factors,

[0023]

[0024] in, and represents the weights at different time points, Indicates the historical transmission failure rate at the current moment, Indicates the historical transmission failure rate at the last moment,

[0025]

[0026] in, represents the standard deviation of the historical transmission failure rate, represents the mean of the historical transmission failure rate,

[0027]

[0028] Among them, α represents the weight factor, Represents the previously computed exponentially decaying weighted average.

[0029] Further, transmitting the upgrade package based on the adjusted slice size, obtaining the slice transmission log, and predicting whether the chip will have an upgrade failure based on the slice transmission log, including:

[0030] Obtain historical shard-level log information of the chip;

[0031] Preprocessing the log information, the preprocessing including: filling missing values, processing outliers and removing duplicates;

[0032] Extracting shard-related features, chip status features, and time-related features from the preprocessed log information, wherein the shard-related features include: shard size, shard sequence number, and shard type (such as code segment, data segment, etc.), and the chip status features include: CPU utilization statistics (mean, maximum, minimum, variance, etc.), memory utilization statistics, and network delay statistics (such as mean, maximum, minimum, variance, etc.);

[0033] The features extracted from the log information are divided into a training set and a test set, the long short-term memory network model is trained using the data in the training set, and the parameters of the long short-term memory network model are adjusted to the optimal value through the test set to obtain the final long short-term memory network model;

[0034] The shard transmission log and chip status data during the chip upgrade process are collected in real time, and the shard transmission log and chip status data are input into the long short-term memory network model for prediction:

[0035] Get the prediction result. If the upgrade is predicted to be abnormal, roll back to the previous stable version and do not upgrade. If the upgrade is predicted to be normal, continue to upgrade.

[0036] Furthermore, the log information includes: fragment sequence number, fragment size, transmission time, fragment retransmission times, chip CPU utilization, chip memory utilization and chip upgrade results.

[0037] Furthermore, a computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the steps of the chip online upgrade method.

[0038] Furthermore, a chip device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the chip online upgrade method.

[0039] The beneficial effects of the present invention are as follows: improving transmission efficiency, maximizing the use of network bandwidth, reducing transmission time, and improving overall transmission efficiency by adjusting the fragment size in real time; reducing the fragment size when the network condition is poor to reduce the retransmission overhead caused by the failure of large fragment transmission, and increasing the fragment size when the network condition is good to reduce the protocol overhead; improving transmission reliability, adjusting the fragment size based on network performance, helping to reduce transmission failures and the number of retransmissions, and improving the transmission success rate; recording and analyzing the fragment transmission log, helping to monitor and identify problems that may occur during the transmission process, taking corrective measures in time, and increasing the reliability of transmission; collecting and analyzing the fragment transmission log in real time, and being able to dynamically monitor the transmission status during the upgrade process; the prediction model can give an upgrade risk assessment based on real-time data, and provide necessary warnings and response measures, such as terminating the upgrade, rolling back to a stable version, etc.; performing fault prediction based on the fragment transmission log, helping to identify potential transmission problems in advance, and reducing the risk of upgrade failure. In high-risk situations, you can choose to roll back to a stable version to ensure system stability. By improving transmission efficiency and reliability, you can reduce interruption time during the upgrade process and improve user experience. Dynamic adjustment strategies and fault prediction mechanisms help provide a more robust and smooth upgrade experience. By establishing a network connection between the chip and the upgrade server, the shard size is adjusted in real time to optimize transmission efficiency, and fault prediction is performed based on the shard transmission log, thereby improving the transmission efficiency, reliability and stability of the upgrade process. It can not only dynamically adapt to different network conditions and reduce the probability of transmission failure, but also detect and respond to potential upgrade failures in advance through prediction mechanisms, ensuring stable system operation and a good user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of a chip online upgrade method according to the present invention. DETAILED DESCRIPTION

[0041] In order to more clearly understand the above-mentioned purpose, features and advantages of the present invention, the present invention is described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0042] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. The embodiments described are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention herein are only for the purpose of describing specific embodiments and are not intended to limit the present invention.

[0044] An embodiment of the present invention provides a chip online upgrade method, the method comprising:

[0045] The chip establishes a network connection with the upgrade server, and the chip sends the current firmware version information to the upgrade server to check whether there is a new update package available;

[0046] If a new update package is available, the update server will fragment the update package and send it to the chip via the network. The update server will adjust the fragment size in real time according to the network performance;

[0047] The upgrade package is transmitted based on the adjusted slice size, a slice transmission log is obtained, and whether the chip will have an upgrade failure is predicted based on the slice transmission log.

[0048] The working principle and effect of the above technical solution are as follows: the chip establishes a preliminary network connection with the upgrade server, the chip sends its current firmware version information to the upgrade server to request an update, the upgrade server checks whether a new update package is available based on the received firmware version information, if a new update package is available, the upgrade server will notify the chip and prepare for the fragmented transmission of the update package, the upgrade server adjusts the size of each fragment in real time according to the current network performance to optimize transmission efficiency and reliability, the update package is divided into multiple small fragments, each fragment is transmitted to the chip through the network in turn, and a transmission log is generated during the transmission of each fragment to record detailed information of the transmission (such as transmission time, whether to retransmit, number of failures, etc.), the chip and the upgrade server collect and store the log information of each fragment transmission for subsequent analysis, and use a machine learning model to predict whether the upgrade process may fail based on the collected fragment transmission log data; if a high risk of failure is predicted, the system chooses to terminate the upgrade and take corresponding remedial measures, such as rolling back to a stable version or adjusting the fragment transmission strategy. Improve transmission efficiency by adjusting the fragment size in real time to maximize the use of network bandwidth, reduce transmission time, and improve overall transmission efficiency; when network conditions are poor, reduce the fragment size to reduce the retransmission overhead caused by large fragment transmission failures, and when network conditions are good, increase the fragment size to reduce protocol overhead; improve transmission reliability, adjust the fragment size based on network performance, which helps reduce transmission failures and retransmission times and improve transmission success rate; record and analyze fragment transmission logs to help monitor and identify problems that may occur during transmission, take corrective measures in a timely manner, and increase transmission reliability; collect and analyze fragment transmission logs in real time to dynamically monitor the transmission status during the upgrade process; the prediction model can give an upgrade risk assessment based on real-time data and provide necessary warnings and response measures, such as terminating the upgrade and rolling back to a stable version; fault prediction based on fragment transmission logs helps identify potential transmission problems in advance and reduce the risk of upgrade failure. In high-risk situations, you can choose to roll back to a stable version to ensure system stability. By improving transmission efficiency and reliability, you can reduce interruption time during the upgrade process and improve user experience. Dynamic adjustment strategies and fault prediction mechanisms help provide a more robust and smooth upgrade experience. By establishing a network connection between the chip and the upgrade server, the shard size is adjusted in real time to optimize transmission efficiency, and fault prediction is performed based on the shard transmission log, thereby improving the transmission efficiency, reliability and stability of the upgrade process. It can not only dynamically adapt to different network conditions and reduce the probability of transmission failure, but also detect and respond to potential upgrade failures in advance through prediction mechanisms, ensuring stable system operation and a good user experience.

[0049] In one embodiment of the present invention, the chip establishes a network connection with an upgrade server, the chip sends current firmware version information to the upgrade server, and checks whether a new update package is available, including:

[0050] The chip prepares current firmware version information, wherein the current firmware version information includes a version number and a release time;

[0051] The chip sends an HTTP request to the upgrade server to obtain the latest firmware version information;

[0052] The chip compares the current version with the latest version information returned by the upgrade server to determine whether an update is required.

[0053] The working principle and effect of the above technical solution are as follows: the chip prepares the currently running firmware version information inside it, and the information includes the firmware version number and the release time of the version; the chip sends an HTTP request to the upgrade server, requesting to obtain the latest firmware version information, and the HTTP request includes the chip's unique identifier and the current firmware version information (version number and release time); the upgrade server returns the latest firmware version information according to the chip's request, including the version number, release time and change log or download link of the upgrade package of the latest version; the chip compares the current firmware version information with the latest firmware version information returned by the upgrade server; if the version number returned by the server is higher than the current version number (or the release time is later than the release time of the current version), the chip recognizes that a new version is available; based on the comparison result, if a new version is found to be available, the chip will decide whether to update the firmware. If no firmware newer than the current version is available, no update is required and the current firmware continues to run. Automatically check for updates. Through automated HTTP requests and version information comparison, the chip can automatically check whether a new firmware version is available without manual intervention. By comparing version numbers and release times, the chip can accurately obtain the latest version of firmware information to avoid missing important updates or repeatedly upgrading and storing existing firmware. By comparing version information, upgrades are only performed when an update is indeed available, avoiding unnecessary frequent updates and saving bandwidth and resources. The automated version check and decision-making process improves the flexibility and reliability of firmware updates and reduces the risk of errors in manual operations. The automated update check and decision-making process reduces the user's manual maintenance work and improves the overall user experience. It ensures that the user's device always runs the latest and most secure firmware version, thereby improving the stability and security of the device.

[0054] In one embodiment of the present invention, the chip establishes a network connection with the upgrade server, and the network connection includes Wi-Fi, mobile network, Ethernet, LPWAN, Bluetooth and Zigbee.

[0055] In one embodiment of the present invention, if a new update package is available, the update server segments the update package and sends the segment to the chip via a network, and the update server adjusts the segment size in real time according to network performance, including:

[0056] If a new update package is available, the upgrade server compares the size of the update package to see if it exceeds a preset threshold, and if the size of the update package exceeds the preset threshold, the update package is fragmented;

[0057] The upgrade server adjusts the slice size according to the network performance through an adaptive slice model.

[0058] The working principle and effect of the above technical solution are as follows: when it is determined that a new update package is available, the upgrade server first checks the size of the update package to determine whether it exceeds a preset threshold; the preset threshold is set based on experience or historical data, and is used to determine whether the update package needs to be fragmented; if the size of the update package exceeds the preset threshold, the upgrade server will fragment the update package; the update package is divided into multiple small fragments to facilitate gradual transmission through the network; the upgrade server monitors network performance in real time and obtains parameters such as transmission rate, delay, and packet loss rate; based on the monitored network performance parameters, an adaptive fragmentation model is used to dynamically adjust the size of each fragment to optimize transmission efficiency and reliability; the upgrade server sends the adjusted fragments to the chip one by one through the network; during the transmission process, the server records the transmission log of each fragment, including transmission time, success or failure, number of retransmissions, etc. Through the adaptive sharding model, the upgrade server can dynamically adjust the shard size according to the real-time network performance to ensure the best transmission efficiency under different network conditions; when the network condition is good, the shard size is increased to improve bandwidth utilization; when the network condition is poor, the shard size is reduced to reduce retransmission overhead; large files are divided into multiple small fragments, which can reduce the probability of transmission failure. Even if a fragment fails to be transmitted, only the fragment needs to be retransmitted instead of the entire update package, which improves the overall reliability of transmission; adaptive sharding adjustment according to the actual network condition helps to improve the transmission success rate in a complex network environment; using the adaptive sharding model to dynamically adjust the transmission strategy enhances the system's flexibility and adaptability to different network environments; real-time monitoring and adjustment reduce the risk of transmission failure and retransmission times, ensuring smooth firmware updates; improving user experience and reducing waiting time, the optimized transmission efficiency can shorten the transmission time of the update package and reduce user waiting time; reliability is enhanced, through optimized transmission strategies and reliability guarantees, the success rate of firmware updates is improved, and the troubles caused by update failures are reduced for users.

[0059] In one embodiment of the present invention, a chip online upgrade method is provided, wherein the adaptive slicing model is:

[0060]

[0061] in, represents the fragment size at the i-th transmission time, represents the instantaneous bit error rate at the i-th moment, represents the instantaneous network jitter at the i-th moment, represents the historical transmission failure rate at the i-th moment, represents the instantaneous equipment load factor at the i-th moment, represents the correction factor of the device resource limit at the i-th moment, B represents the base shard size, represents the maximum tolerable bit error rate, Indicates the maximum tolerable network jitter, The length of the time window representing the historical transmission failure rate, Indicates the network bandwidth. Indicates the start time of the time window. Indicates the end time of the time window.

[0062] The device load factor can be calculated by weighted average of the chip CPU, memory and storage usage.

[0063] ,in, Indicates the memory usage of the chip at the current moment. Indicates the CPU usage of the chip at the current moment.

[0064] In one embodiment of the present invention, a chip online upgrade method is provided, wherein the time window length of the historical transmission failure rate is calculated by a time window model. Specifically, the time window model is:

[0065]

[0066] in, represents the weighted difference between the current and past historical transmission failure rates, CV(H) represents the coefficient of variation of the historical transmission failure rate, represents the exponentially decaying weighted average of the historical transmission failure rate, Indicates the preset minimum time window length, Indicates the preset maximum time window length, represents the basic time window length correction factor, , and Represents the power exponent for adjusting the weights of different factors,

[0067]

[0068] in, and represents the weights at different time points, Indicates the historical transmission failure rate at the current moment, Indicates the historical transmission failure rate at the last moment,

[0069]

[0070] in, represents the standard deviation of the historical transmission failure rate, represents the mean of the historical transmission failure rate,

[0071]

[0072] Among them, α represents the weight factor, Represents the previously computed exponentially decaying weighted average.

[0073] In one embodiment of the present invention, an upgrade package is transmitted based on the adjusted slice size, a slice transmission log is obtained, and whether the chip will have an upgrade failure is predicted based on the slice transmission log, including:

[0074] Obtain historical shard-level log information of the chip;

[0075] Preprocessing the log information, the preprocessing including: filling missing values, processing outliers and removing duplicates;

[0076] Extracting shard-related features, chip status features, and time-related features from the preprocessed log information, wherein the shard-related features include: shard size, shard sequence number, and shard type (such as code segment, data segment, etc.), and the chip status features include: CPU utilization statistics (mean, maximum, minimum, variance, etc.), memory utilization statistics, and network delay statistics (such as mean, maximum, minimum, variance, etc.);

[0077] The features extracted from the log information are divided into a training set and a test set, the long short-term memory network model is trained using the data in the training set, and the parameters of the long short-term memory network model are adjusted to the optimal value through the test set to obtain the final long short-term memory network model;

[0078] The shard transmission log and chip status data during the chip upgrade process are collected in real time, and the shard transmission log and chip status data are input into the long short-term memory network model for prediction:

[0079] Get the prediction result. If the upgrade is predicted to be abnormal, roll back to the previous stable version and do not upgrade. If the upgrade is predicted to be normal, continue to upgrade.

[0080] An embodiment of the present invention is a chip online upgrade method, wherein the log information includes: fragment sequence number, fragment size, transmission time, number of fragment retransmissions, chip CPU utilization, chip memory utilization and chip upgrade results.

[0081] The working principle and effect of the above technical solution are as follows: to obtain historical data, first, to extract log information related to shard transmission from the historical records of the chip, which includes the size, serial number, type of the shard, and other status data recorded during the transmission process; data preprocessing, filling missing values, interpolating missing records, and using the most recent data or statistical mean and other methods to fill missing data; processing outliers, detecting and removing obviously illogical data, such as extreme values ​​or erroneous data, to ensure data quality; deduplication, if there are duplicate records in the log, they will be deleted to avoid interference with subsequent analysis; feature extraction, extracting shard-related features (such as shard size, shard serial number, shard type), chip status features (such as CPU utilization, memory utilization, mean and variance of network delay, etc.) and time-related features (such as timestamp) from the preprocessed log information; dividing the extracted feature data set into training set and test set, usually at 70%-80 % is used for training, and the rest is used for testing; the training set is used to train the long short-term memory network (LSTM), which can process time series data and capture the temporal relationship in the data; during the training process, the model parameters are continuously adjusted through the back propagation algorithm until the prediction performance of the model reaches the expected level; the performance of the model is verified using the test set, and the effect of the model is evaluated by calculating various evaluation indicators (such as accuracy, F1 value, etc.), and the parameters are adjusted to ensure the ultimate stability and accuracy of the model; during the chip upgrade process, the shard transmission log and chip status data are collected in real time, and these data are input into the trained LSTM model; the model outputs the prediction result to determine whether anomalies may occur during the upgrade process; if the LSTM model predicts that anomalies may occur, the fallback mechanism is executed to restore to the previous stable version to avoid unnecessary upgrade risks; if the prediction result shows normal, the upgrade operation continues to ensure the smooth progress of the firmware update. Improve the security of the upgrade. By using the LSTM model to analyze historical data, it can predict in advance the anomalies that may be encountered during the upgrade, thereby greatly reducing the risk of failure. It can make decisions in real time based on the chip status and shard transmission status, has strong adaptability, and can cope with different network and device conditions. The automated prediction and fallback mechanism reduces the need for manual user intervention and improves the intelligence level of the system. Improve user experience by avoiding device unavailability due to upgrade failure, and enhance user trust and satisfaction with the device upgrade process. Efficient resource utilization. Through real-time monitoring and prediction, the system can optimize its resource utilization when necessary to ensure that the upgrade can be effectively implemented under resource constraints. It integrates data collection, preprocessing, feature extraction, model training, and real-time monitoring, and uses long short-term memory networks (LSTM) to make intelligent predictions and decisions on the firmware upgrade process.Its working principle ensures real-time response and reduction of potential risks in complex and changing environments, providing a reliable solution for online chip upgrades. It is especially suitable for the Internet of Things, smart homes, industrial control and other fields, and can effectively improve the upgrade efficiency and reliability of equipment.

[0082] An embodiment of the present invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, the processor executes the steps of the chip online upgrade method.

[0083] According to one embodiment of the present invention, a chip device includes a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the chip online upgrade method.

[0084] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

Claims

1. A chip online upgrade method, characterized in that: The method comprises: The chip establishes a network connection with the upgrade server, and the chip sends the current firmware version information to the upgrade server to check whether there is a new update package available; If a new update package is available, the update server will fragment the update package and send it to the chip through the network. The update server adjusts the fragment size in real time according to the network performance, including: If a new update package is available, the upgrade server compares the size of the update package to see if it exceeds a preset threshold, and if the size of the update package exceeds the preset threshold, the update package is fragmented; The upgrade server adjusts the slice size through an adaptive slice model according to network performance. Specifically, the adaptive slice model is: in, represents the fragment size at the i-th transmission time, represents the instantaneous bit error rate at the i-th moment, represents the instantaneous network jitter at the i-th moment, represents the historical transmission failure rate at the i-th moment, represents the instantaneous equipment load factor at the i-th moment, represents the correction factor of the device resource limit at the i-th moment, B represents the base shard size, represents the maximum tolerable bit error rate, Indicates the maximum tolerable network jitter, The length of the time window representing the historical transmission failure rate, represents the network bandwidth at the i-th transmission moment, which varies with time t, where t represents the time variable. Indicates the total bandwidth in this time window, Indicates the start time of the time window. Indicates the end time of the time window; The upgrade package is transmitted based on the adjusted slice size, a slice transmission log is obtained, and whether the chip will have an upgrade failure is predicted based on the slice transmission log.

2. According to claim 1, a chip online upgrade method is characterized in that: The chip establishes a network connection with the upgrade server, sends the current firmware version information to the upgrade server, and checks whether a new update package is available, including: The chip prepares current firmware version information, wherein the current firmware version information includes a version number and a release time; The chip sends an HTTP request to the upgrade server to obtain the latest firmware version information; The chip compares the current version with the latest version information returned by the upgrade server to determine whether an update is required.

3. According to claim 1, a chip online upgrade method is characterized in that: The chip establishes a network connection with the upgrade server, and the network connection includes Wi-Fi, mobile network, Ethernet, LPWAN, Bluetooth and Zigbee.

4. A chip online upgrade method according to claim 1, characterized in that: The time window length of the historical transmission failure rate is calculated by a time window model. Specifically, the time window model is: in, represents the weighted difference between the current and past historical transmission failure rates, CV(H) represents the coefficient of variation of the historical transmission failure rate, represents the exponentially decaying weighted average of the historical transmission failure rate, Indicates the preset minimum time window length, Indicates the preset maximum time window length, represents the basic time window length correction factor, , and Represents the power exponent for adjusting the weights of different factors, in, and represents the weights at different time points, Indicates the historical transmission failure rate at the current moment, Indicates the historical transmission failure rate at the last moment, in, represents the standard deviation of the historical transmission failure rate, represents the mean of the historical transmission failure rate, Among them, α represents the weight factor, Represents the previously computed exponentially decaying weighted average.

5. According to claim 1, a chip online upgrade method is characterized in that: Transmitting the upgrade package based on the adjusted slice size, obtaining the slice transmission log, and predicting whether the chip will have an upgrade failure based on the slice transmission log, including: Obtain historical shard-level log information of the chip; Preprocessing the log information, the preprocessing including: filling missing values, processing outliers and removing duplicates; Extracting shard-related features, chip status features, and time-related features from the preprocessed log information, wherein the shard-related features include shard size, shard sequence number, and shard type, and the chip status features include CPU utilization statistics, memory utilization statistics, and network delay statistics; The features extracted from the log information are divided into a training set and a test set, the long short-term memory network model is trained using the data in the training set, and the parameters of the long short-term memory network model are adjusted to the optimal value through the test set to obtain the final long short-term memory network model; The shard transmission log and chip status data during the chip upgrade process are collected in real time, and the shard transmission log and chip status data are input into the long short-term memory network model for prediction: Get the prediction result. If the upgrade is predicted to be abnormal, roll back to the previous stable version and do not upgrade. If the upgrade is predicted to be normal, continue to upgrade.

6. A chip online upgrade method according to claim 5, characterized in that: The log information includes: fragment sequence number, fragment size, transmission time, fragment retransmission times, chip CPU utilization, chip memory utilization and chip upgrade results.

7. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 6.

8. A chip device, comprising a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 6.

Images