Firmware update processing method and device based on AI drive, terminal and medium

Abstract

The invention discloses a firmware update processing method and device based on AI drive, a terminal and a medium, and belongs to the technical field of intelligent equipment firmware updation.The method comprises the steps that log data and sensor data of an intelligent terminal are obtained, and user use habit data are analyzed through the log data and the sensor data; according to the analyzed use habit data of the user, predicting a future idle time period of the corresponding intelligent terminal by adopting an AI prediction model of a time sequence; according to the predicted future idle time period of the corresponding intelligent terminal, selecting an updating opportunity for confirming firmware updating; and based on the update opportunity of the selected and confirmed firmware update, the arrival of the corresponding opportunity is detected, and the firmware update is controlled to be automatically executed. The method has the advantages of avoiding triggering update when the user uses the equipment at high frequency, improving the update success rate and reducing the system resource consumption, and provides convenience for the user to use.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent device firmware update, in particular to a firmware update processing method and device based on AI driving, an intelligent terminal and a storage medium. BACKGROUND

[0002] With the continuous upgrading of the functions of intelligent devices such as smart TVs, firmware update has become an important means to improve the performance and security of the device. However, the traditional firmware update strategy in the prior art mostly adopts a fixed time or manual triggering mode, which is likely to cause interference (such as interrupting the viewing experience) when the user uses the device. In the prior art, although some solutions attempt to update according to the idle time of the device, they lack in-depth analysis of the user's usage habits, resulting in inaccurate selection of update timing.

[0003] Therefore, the prior art still needs to be improved and developed. SUMMARY

[0004] To solve the above technical problems, the present application provides a firmware update processing method and device based on AI driving, an intelligent terminal and a storage medium, which solves the problem of lack of in-depth analysis of user usage habits in the prior art firmware update, resulting in inaccurate selection of update timing; has the advantages of avoiding triggering updates when the user frequently uses the device, improving the success rate of updates, and reducing system resource consumption.

[0005] A firmware update processing method based on AI driving, comprising: Obtaining log data and sensor data of an intelligent terminal, and analyzing user usage habit data from the log data and sensor data; According to the analyzed user usage habit data, an AI prediction model of time series is used to predict the future idle time period of the corresponding intelligent terminal; According to the predicted future idle time period of the corresponding intelligent terminal, the update timing of the confirmed firmware update is selected; Based on the selected update timing of the confirmed firmware update, when the corresponding timing arrives, the firmware update is automatically executed.

[0006] The firmware update processing method based on AI driving, wherein after the step of detecting the arrival of the corresponding timing and controlling the automatic execution of the firmware update based on the selected update timing of the confirmed firmware update, the method further comprises: Receiving user feedback information on the firmware update, and optimizing the AI prediction model based on the feedback information and the firmware update result.

[0007] The AI-driven firmware update processing method, wherein the step of obtaining log data and sensor data of the intelligent terminal, and analyzing user usage habit data based on the log data and sensor data comprises: Obtaining log data and sensor data of the intelligent terminal, wherein the log data comprises: viewing duration data, usage frequency data, and device operation type data; and the sensor data comprises: camera detection data of whether a user is in front of the device; Using a clustering algorithm to analyze user behavior patterns based on the log data and sensor data, and using a statistical method to calculate user usage habits, wherein the usage habits comprise: average viewing duration and peak time period; Outputting a user behavior feature vector as user usage habit data according to the analyzed user behavior patterns and usage habits.

[0008] The AI-driven firmware update processing method, wherein the step of predicting a future idle time period of the corresponding intelligent terminal based on the analyzed user usage habit data using a time series AI prediction model comprises: Inputting the user behavior feature vector as user usage habit data into the time series AI prediction model, and outputting a probability distribution of the future idle time period; Wherein the user behavior feature vector comprises: user historical usage data, date type of weekdays and weekends, and legal holiday information; According to the output probability distribution of the future idle time period, selecting and optimizing model parameters through cross-validation to obtain the predicted future idle time period of the corresponding intelligent terminal.

[0009] The AI-driven firmware update processing method, wherein the step of selecting and confirming an update time of the firmware update based on the predicted future idle time period of the corresponding intelligent terminal comprises: Obtaining the predicted future idle time period of the corresponding intelligent terminal, and selecting a time period with an idle probability reaching a predetermined value and a long duration as the update time of the firmware update; Wherein the update time of the firmware update is an update time including a specific time point.

[0010] The AI-driven firmware update processing method, wherein the step of detecting the arrival of the corresponding time based on the selected and confirmed update time of the firmware update, and controlling automatic execution of the firmware update comprises: Controlling to download a firmware update package in advance in a space-time period, and verifying the integrity of the update package; Based on the selected and confirmed update time of the firmware update, analyzing the specific time point of the update time; When the specific time point of the update opportunity is detected, the control automatically performs the firmware update on the downloaded firmware update package; When the firmware update package update installation is completed, the control restarts the intelligent terminal.

[0011] The AI-driven firmware update processing method, wherein the update opportunity of the selected and confirmed firmware update includes the step of: The firmware update process is pre-set to support breakpoint resume, which is used to avoid update failure caused by network interruption; The firmware update process is pre-set to support rollback mechanism, which is used to ensure that the original version can be restored after the update fails; The firmware update process is pre-set to save log records at any time, which is used to record the update process and results.

[0012] An AI-driven firmware update processing device, wherein the device comprises: A user behavior analysis module is used to obtain log data and sensor data of the intelligent terminal, and analyze user usage habit data through the log data and sensor data; An idle time prediction module is used to predict the future idle time period of the corresponding intelligent terminal by using the time sequence AI prediction model according to the analyzed user usage habit data; An update opportunity optimization module is used to select and confirm the update opportunity of the firmware update according to the predicted future idle time period of the corresponding intelligent terminal; An update execution module is used to detect the arrival of the corresponding time based on the selected and confirmed update opportunity of the firmware update, and control the automatic execution of the firmware update; A feedback optimization module is used to receive feedback information of the firmware update from the user, and optimize the AI prediction model through the feedback information and the firmware update result.

[0013] An intelligent terminal, comprising a memory and one or more programs, wherein one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs include a program for executing any one of the methods.

[0014] A computer readable storage medium, wherein when the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute any one of the methods.

[0015] From the above, the application provides an AI-driven firmware update processing method and device, an intelligent terminal and a storage medium, which automatically selects the best update time to perform firmware update by analyzing user usage habit data and predicting future idle time periods, has the advantages of avoiding triggering update when the user frequently uses the device, improving update success rate and reducing system resource consumption.

[0016] And by analyzing user usage habits to select the best time for firmware update, the user's interference is reduced, and the user's use is facilitated. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0018] Figure 1 is a flowchart of the AI-driven firmware update processing method provided by the embodiment 1 of the present application.

[0019] Figure 2 The principle block diagram of the AI-driven firmware update processing device embodiment provided by the present application.

[0020] Figure 3 is the internal structure principle block diagram of the intelligent terminal provided by the embodiment of the present application. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical scheme and advantages of the present application more clear and definite, the following will further describe the present application in detail with reference to the drawings and examples. It should be understood that the specific examples described herein are only used to explain the present application, and are not used to limit the present application.

[0022] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement condition, etc. between the components in a certain specific posture (as shown in the drawings), if the specific posture changes, the directional indications will also change accordingly.

[0023] With the continuous upgrading and iteration of the functions of intelligent terminal devices, firmware updating has become an important means to improve device performance, repair system vulnerabilities, and enhance security protection. However, the traditional firmware updating strategy in the prior art has significant defects: on the one hand, the batch updating method with a fixed time period often triggers the updating process when the user frequently uses the device, causing interruption of device use and seriously affecting user experience, for example, suddenly popping up an update prompt when the user is watching a video or performing an important operation; on the other hand, the update mode completely dependent on user manual triggering is prone to update lag due to user forgetfulness or procrastination, leaving the device in a long-term security risk. Although some improved solutions attempt to perform updates by detecting the current idle state of the device, such methods only focus on the instantaneous state and lack deep learning of user long-term usage habits, and cannot accurately predict the future optimal update window. More critically, the prior art fails to effectively integrate multi-dimensional information such as device log data and sensor data, resulting in significant errors in judging user behavior patterns, and often causing the predicted idle time period to mismatch the actual usage demand. Such a rough update timing selection mechanism not only reduces the success rate of updating, but also may cause repeated attempts to update and exacerbate system resource consumption. In addition, existing solutions generally lack a closed-loop optimization mechanism, and cannot continuously improve the prediction model according to the actual update effect, making it difficult for the system to adapt to changes in user behavior patterns.

[0024] To solve the above problems, the present inventors have observed that the behavior of users operating devices has predictable periodic characteristics. By analyzing the interaction time points recorded in the device log and the environmental state data collected by the sensor, a behavior model reflecting the user's daily usage habits can be constructed. Based on this model, the time interval in which the device is likely to be in a non-use state in the future can be further inferred, providing accurate time window prediction for firmware updating. This approach breaks through the limitations of traditional solutions that only focus on the current device state, and changes passive response to active prediction, fundamentally solving the update interference problem.

[0025] Embodiment one As shown in Figure 1 The AI-driven firmware updating processing method of the embodiment one of the present application comprises the following steps: Step S100, obtaining log data and sensor data of the intelligent terminal, and analyzing user usage habit data through the log data and sensor data; In the embodiment of the present step, in the data acquisition link, log data and sensor data of the intelligent terminal are acquired, wherein the intelligent terminal includes intelligent television, smart phone, smart watch, smart home device and various devices with data recording and sensing functions. The log data is automatically generated by the intelligent terminal during operation, such as application startup time, use time, page jump path of the mobile phone, switch time and operation mode adjustment record of the smart home device, which records the operation track and state change of the device in detail. The sensor data is collected by various sensors built-in the terminal, such as the acceleration sensor of the mobile phone can reflect the motion state of the user, the light sensor can record the change of the ambient light, and the heart rate sensor of the smart watch can monitor the physiological indicators of the user, and these data can capture the behavior and environment characteristics of the user from the objective point of view. In the embodiment of the present step, in the data analysis link, the log data and sensor data of the intelligent terminal are acquired, and statistical methods and data mining methods are used for in-depth processing. Through the analysis of the application use frequency and time period in the log data, the preferences of the user for different functions can be summarized; combined with the association between the motion data and the application use in the sensor data, the use habits of the user in a specific motion state can be found, and finally the use habit data reflecting the behavior mode, preference tendency of the user can be extracted.

[0026] The user use habit data analyzed in the embodiment can provide a strong reference for the firmware update and optimization of the intelligent terminal product.

[0027] Step S200, according to the analyzed user use habit data, a time series AI prediction model is used to predict the future idle time period of the corresponding intelligent terminal; In the embodiment of the present step, on the basis of the foregoing acquisition and analysis of the user use habit data, the idle time of the terminal is further predicted by means of a specific AI model, which mainly includes: In the embodiment, a time series AI prediction model tool is used to process the data with time sequence in the user use habit data, which can capture the law and trend of data change with time, such as the use law of the intelligent terminal by the user on different dates and at different time periods. The integration of AI technology makes the model have stronger learning ability, which can mine the hidden patterns from the massive user use habit data, such as the use time difference between weekdays and weekends, special use law during holidays, etc., thereby improving the accuracy of the prediction. In the embodiment of the present application, the prediction process is closely related to the user usage habit data. These data record the time distribution of the user's past use of the terminal, such as which time period is frequently used and which time period is almost not used every day. The model learns the time characteristics in these historical data, combines the change trend of time sequence, and calculates the time period in which the user is likely not to use the terminal in the future, i.e. the future idle time period of the intelligent terminal. For example, if the data shows that the user almost does not use the mobile phone from 11pm to 7am the next day, the model will predict that this time period of the following day is the idle time period of the mobile phone based on this rule. In this way, the present application can intelligently process the firmware upgrade according to the idle time of the intelligent terminal.

[0028] Step S300, selecting an update time for confirming the firmware update according to the predicted future idle time period of the corresponding intelligent terminal; In the embodiment of this step, it is the actual application link after the prediction of the future idle time period of the intelligent terminal, and the core is to combine the prediction result with the firmware update demand to accurately determine the time point of the update execution.

[0029] Specifically, the firmware update usually needs the intelligent terminal to invest certain computing resources and network traffic, and may temporarily affect the normal use of the terminal during the process. Therefore, in the embodiment of this step, the most suitable time period for firmware update is selected from the predicted multiple future idle time periods. For example, if it is predicted that a smart watch is idle from 2am to 4am every day, and the network environment is stable and the terminal has sufficient power during this time period, then this time period will be determined as the time for firmware update; if there are poor network signal or low power state in the predicted idle time period, these time periods will be excluded and a better time will be selected. At the same time, some terminals will send a prompt to the user after determining the update time, and the update will be executed after the user confirms, so as to further protect the user's right to use. In this way, the present application has the following advantages: 1) Reducing the interference of the update on the user: If the firmware update is performed when the user uses the terminal, it may cause problems such as operation interruption and application crash. However, if the update is performed in the predicted idle time period, such interference can be avoided to the greatest extent, ensuring that the user obtains a smooth experience during the peak time period of using the terminal and reducing user complaints caused by the update. 2) Improving the success rate of firmware update: In the idle time period, the terminal is usually not frequently operated, and the network resources are relatively sufficient, which can reduce the update failure caused by user misoperation, network fluctuation and other factors. In addition, some terminals will automatically ensure that the power is within a reasonable range during the idle time period, avoiding interruption of the update due to insufficient power, and further improving the success rate of the update. 3) Reduce the occupation conflict of terminal resources: the firmware update process will occupy certain CPU, memory and other resources. If it is carried out when the user uses the terminal, it may compete with the user's running application for resources, causing the terminal to lag. Selecting the idle period for updating can make the terminal focus on completing the update task, reduce resource occupation conflicts, and ensure the stable running performance of the terminal before and after updating. 4) Optimize the utilization efficiency of network bandwidth: for firmware updates that require a large amount of network data transmission, performing in the idle time period can avoid network usage peaks and slow update speeds caused by bandwidth congestion. This not only speeds up the update progress of individual terminals, but also reduces the occupancy pressure on the overall network bandwidth, especially suitable for scenarios where multiple terminals are updating simultaneously.

[0030] Step S400, based on the selected and confirmed firmware update time, detects that the corresponding time has arrived, and controls automatic execution of firmware update.

[0031] This step is the final execution link of the firmware update process, which connects the update time selected and confirmed in the previous steps, and realizes the non-sensing update of firmware through an automatic mechanism, mainly including three core actions. First, based on the selected and confirmed update time, it means that the entire execution process is based on the idle time period determined earlier, which conforms to the user's usage habits and terminal state, such as the confirmed time period of 2:00 to 4:00, to ensure that the update does not deviate from the pre-set reasonable range. Second, detecting that the corresponding time has arrived, the intelligent terminal will track the current time in real time through the built-in time monitoring module. When the system detects that the time enters the pre-set update period, it will trigger the subsequent update instruction. At the same time, some terminals will also quickly check the terminal state again when the time arrives, such as confirming whether the network is still stable, whether the power is sufficient, etc., to deal with possible sudden conditions and further ensure the effectiveness of the update conditions. Finally, control the automatic execution of firmware update. When the time and state are both satisfied, the control system of the terminal will automatically start the firmware update program, including downloading the update package from the server, verifying the file integrity, performing system upgrade operations, etc. The entire process does not require manual intervention, achieving fully automated operation. In this way, the present application has the following advantages: 1) Full-process automation of firmware update can be achieved: from the selection and confirmation of the update time to the final execution, forming a complete automatic closed loop, reducing the involvement of manual operation. This not only reduces the update delay problem caused by user forgetfulness or operational errors, but also improves the execution efficiency of firmware update, ensuring that the terminal can obtain functional optimization and security patches in a timely manner. 2) Ensure the stability and reliability of the update process: secondary state detection before automatic execution can effectively avoid update failure caused by sudden conditions (such as network interruption, power drop). At the same time, the automatic update process follows the preset standard steps, reducing the risk of misoperation caused by human operation, making the update process more stable and reliable. 3) Can improve the degree of non-sensitization of user experience: since the update is automatically executed in the preset idle period, the user will not feel the existence of the update process in daily use of the terminal, avoiding the waiting, operation interruption and other situations that may occur when manually triggering the update. This non-sensitization experience can make the user more focused on the use of the terminal, enhancing the user's satisfaction with the product. From the above, the technical scheme of obtaining log data and sensor data of the intelligent terminal, obtaining user usage habit data by analyzing the data, using a time series prediction model to infer the future idle time period of the device, selecting the best update time according to the prediction result, and automatically executing the firmware update when the time is detected to arrive is proposed in the above embodiments of the present application.

[0032] Among them, the log data refers to the interaction event information automatically recorded during the operation of the device, which can be realized by a system log collection module, for example, recording the time when the user starts the device each time, the type and duration of the operation application. The sensor data refers to the monitoring data reflecting the user's physical environment state, which can be realized by a camera or an infrared sensor, for example, judging whether the user is in front of the device through image recognition. The time series prediction model refers to a machine learning algorithm for processing time-related data, which can be realized by a long short-term memory neural network. The model can capture the periodicity of user behavior in the time dimension. The idle time period selection refers to selecting the time period that meets the update operation requirements from the prediction result, which can be realized by setting the probability threshold and the minimum duration condition, for example, only selecting the time period with an idle probability of more than 80% and a continuous duration of more than 30 minutes.

[0033] Specifically, the system first extracts historical log data from the device storage unit, and reads real-time picture data collected by the camera sensor at the same time. The data processing module cleans and extracts features from the original data, for example, converts the start and end time of the user's daily use of the device into a time series format. The trained time series model receives the processed data and outputs the probability distribution of the device being in idle state in each time period within the next 24 hours. The decision module selects the optimal update period according to the preset strategy, for example, preferentially selecting the time period in the late night and the time period with stable device network connection. The update execution module automatically starts the download and installation process when the predetermined time arrives, and restarts the device after completion. The user does not need to perform any manual operation during the whole process, and the system completes the firmware upgrade in a background silent mode.

[0034] Compared with the prior art, the traditional scheme relies on a static schedule or instant state detection, and is prone to interruption of updating due to temporary use of the device by the user. The scheme dynamically predicts the user behavior pattern through a machine learning model, and can plan the update task in advance without the user's awareness. For example, the prior art may misjudge a short absence of the user from the device as an available period, while the scheme can accurately distinguish between temporary absence and long absence of the user by combining long-term use rules and real-time sensor data.

[0035] Through the above technical scheme, the present application realizes adaptive matching of the firmware update timing and the user behavior pattern, effectively avoiding interference of the updating process on normal use of the device. The system can dynamically adjust the updating strategy according to the use characteristics of individual users, for example, automatically postponing the updating to the early morning period for a user who frequently uses the device at night. This intelligent scheduling mechanism based on behavior prediction significantly improves the success rate of firmware updating and user satisfaction.

[0036] The present application further proposes receiving feedback information of the user on the firmware updating, and optimizing a prediction model of an AI artificial intelligence model through the feedback information and a firmware updating result.

[0037] The feedback information refers to evaluation data of the user on the running state of the device after updating, and can be implemented by using score data actively submitted by the user or operation behavior data automatically collected by the system, and is used to reflect the satisfaction of the user on the updating effect. The firmware updating result refers to installation success rate, updating time consumption and device performance change data recorded in the updating process, and can be obtained through a system log collection and performance monitoring module, and is used to quantify the actual influence of the updating operation. The optimized prediction model refers to adjusting model parameters based on a supervised learning framework, and can be implemented by using a gradient descent algorithm or a reinforcement learning method to iteratively train a time series prediction model, so that the model can dynamically correct the idle period prediction logic according to historical feedback data.

[0038] Specifically, after automatically performing the firmware updating, the intelligent terminal continuously collects score data of the user on the device response speed, interface smoothness and other dimensions, and records whether an abnormal interruption or version rollback event occurs in the updating process. After desensitization processing, these data form a training sample set with original prediction data, the deviation value of the prediction period selection and the actual perception of the user is calculated through a loss function, and then the weight parameters of the hidden layer in the LSTM neural network are adjusted in a reverse direction. For example, when the user feedback indicates that the device is stuck due to the selection of the updating period, the model will reduce the updating decision weight under similar idle probability threshold conditions.

[0039] Compared with the prior art, the traditional scheme only selects the update period according to the preset rule and lacks a dynamic optimization mechanism, and the application constructs a closed-loop feedback system, taking the user subjective experience and objective update result as the model optimization basis. The method of fixed threshold value in the prior art cannot adapt to the behavior changes of different user groups, and the application can track the user habit migration in real time through the online learning mechanism.

[0040] Through the above technical scheme, the application realizes the adaptive evolution of the prediction model, effectively solves the update period misjudgment problem caused by the change of user behavior mode. By fusing user feedback and device state data for training, the idle period prediction accuracy is continuously improved, thereby reducing the interference probability of firmware update on normal use of users, and reducing the device abnormal risk caused by update failure.

[0041] The application further proposes to obtain log data and sensor data of the intelligent terminal, analyze user usage habit data through the log data and sensor data, including: obtaining log data and sensor data of the intelligent terminal, wherein the log data includes viewing time data, usage frequency data, and device operation type data; the sensor data includes data detected by a camera whether the user is in front of the device; using a clustering algorithm to analyze the user behavior mode, and using a statistical method to calculate the user usage habit, wherein the usage habit includes average viewing time and peak period; outputting the user behavior feature vector as the user usage habit data according to the analyzed user behavior mode and usage habit.

[0042] The log data refers to structured data recording user operation behavior, which can be implemented by a system log collection module, for example, the time stamp and operation type of each user operation are captured by a burying point technology. The sensor data refers to environmental state information obtained by a hardware sensor, which can be implemented by a camera combined with a face recognition algorithm, for example, whether the user is in front of the device is detected by an infrared sensor. The clustering algorithm refers to an unsupervised learning method for discovering the internal distribution law of data, which can be implemented by a K-means algorithm, for example, user operation behavior is divided into different categories according to time distribution to identify the behavior mode. The statistical method refers to a data analysis means based on mathematical operation, which can be implemented by a sliding window mean calculation, for example, the average active time of the user in a specific time period is calculated.

[0043] Specifically, the data collection module continuously collects operation logs and sensor signals of the intelligent terminal, wherein the log data is classified into dimensions such as viewing duration and operation frequency, and the sensor data is converted into a binary state identifier. The data preprocessing stage inputs the standardized original data into a clustering model, for example, maps the user's daily operation time points into two-dimensional coordinates for clustering, and identifies high-frequency operation periods and low-frequency operation periods. The statistical module performs secondary analysis on the clustering results, for example, calculates the peak period difference between weekdays and weekends, and finally generates a user behavior vector containing time distribution characteristics. The vector is encapsulated into a data structure containing multiple features, for example, the user's historical active period, device idle regularity and other information are stored in JSON format.

[0044] Compared with the prior art, the traditional scheme only relies on simple time period statistics or device idle state detection, for example, only records the device boot-up duration as an update basis. However, the present scheme can eliminate the misjudgment of the device standby but the user still present by fusing multi-source data to construct a behavior model, for example, combining camera detection results and operation logs to cross-verify the user's real use state. In addition, the cooperative application of clustering algorithm and statistical method can identify non-explicit behavior rules, for example, find the user's habit of briefly waking up the device at night, thereby avoiding triggering updates during non-continuous idle periods.

[0045] Through the above technical scheme, the present application can accurately identify the spatio-temporal rules of the user's real use of the device, and avoid the update interference problem caused by the device standby but the user present. The user habit data generated based on the behavior feature vector provides high-precision input for subsequent prediction models, for example, by distinguishing the use difference between weekdays and holidays, improving the accuracy of idle time period prediction, and finally realizing the dynamic optimization of the firmware update timing.

[0046] The present application further proposes a step of predicting the future idle time period of the corresponding intelligent terminal according to the analyzed user use habit data, which includes inputting the user behavior feature vector as the user use habit data into the AI prediction model adopting time series, outputting the probability distribution of the future idle time period, wherein the user behavior feature vector includes user historical use data, date type of weekdays and weekends, and legal holiday information, according to the output probability distribution of the future idle time period, selecting and optimizing the model parameters through cross-validation, and obtaining the predicted future idle time period of the corresponding intelligent terminal.

[0047] The user behavior feature vector refers to integrating user historical use data, date type and holiday information into a multi-dimensional data set. Specifically, a data normalization method can be used to convert data of different dimensions into a unified numerical range, thereby providing standardized input for the model. The AI prediction model of time series refers to a machine learning model that predicts future states based on historical data sequences. Specifically, a long short-term memory network or a Transformer architecture can be used to capture the time dependence of user behavior. The probability distribution refers to the quantitative result of the possibility of the idle time period output by the model. Specifically, a probability density function or a confidence interval can be used to represent the selectability of different time periods. Cross-validation refers to a method of optimizing model parameters by dividing the training set and the validation set. Specifically, K-fold cross-validation can be used to adjust model hyperparameters, thereby avoiding overfitting and improving prediction stability.

[0048] Specifically, after the user behavior feature vector is input into the time series prediction model, the model generates an idle probability distribution for multiple future time periods by analyzing historical use patterns and date attributes. For example, the model can identify that the user has a low use frequency in the evening on weekends, and predict a continuous idle period in combination with holiday information. Subsequently, the model parameters are iteratively optimized through cross-validation, such as adjusting the number of neural network layers or the learning rate, and finally the parameter combination with the smallest prediction error is selected to determine the optimal future idle time period prediction result.

[0049] Compared with the prior art, the existing scheme usually only relies on the current state of the device or simple historical records to determine the idle time, and does not consider the influence of date type, holiday and other external factors on user behavior. The present scheme constructs a user behavior vector by fusing multi-dimensional features, captures long-term patterns using a time series model, and improves prediction accuracy using cross-validation, making the idle time period prediction more consistent with the actual use scenario of the user.

[0050] Through the above technical solutions, the present application can dynamically predict the idle time period based on user historical behavior patterns and date attributes, effectively avoiding prediction bias caused by ignoring holidays or workdays. At the same time, the cross-validation mechanism enhances the generalization ability of the model, ensuring the reliability of the prediction result in different scenarios, thereby providing more accurate data support for the selection of firmware update timing.

[0051] The present application further proposes an AI-driven firmware update processing method, which includes selecting an update timing for firmware update according to the predicted future idle time period of the intelligent terminal, including obtaining the predicted future idle time period, and selecting a time period with an idle probability reaching a predetermined value and a long duration as the update timing for firmware update, wherein the update timing for firmware update is an update timing including a specific time point.

[0052] The future idle time period refers to a time period in which the intelligent terminal is not used, which can be achieved by using a time series prediction model to calculate the probability distribution of user historical behavior data. The idle probability reaching a predetermined value refers to the probability of the idle time period output by the model exceeding a set threshold, which can be achieved by using a probability threshold screening mechanism. The long duration refers to the length of the predicted idle time period meeting the minimum time requirement of the firmware update, which can be achieved by setting a lower limit of the time length. The specific time point refers to the start and end time markers of the time period to the minute level, which can be achieved by using a timestamp analysis algorithm.

[0053] Specifically, after obtaining the future idle time period predicted by the AI model, the system first filters out candidate time periods with an idle probability exceeding a preset threshold (e.g., 85%), and then filters out time periods with a duration meeting the requirements according to the estimated installation time of the firmware update package (e.g., 30 minutes). When there are multiple time periods meeting the conditions, the time period closest to the current time and with the longest duration is selected as the update time. For example, if the prediction shows that the idle probability of 1:00-3:00 the next day is 92% and the duration is 120 minutes, then this time period will be selected as the time window for performing the firmware update.

[0054] Compared with the prior art, the traditional method only performs updates according to the current idle state of the device or fixed maintenance windows, which is prone to interruption due to temporary use of the device by the user. However, the present scheme quantitatively evaluates the idle probability and duration, which can actively avoid high-risk time periods when the user may suddenly use the device, such as in the predicted high-probability idle time period, the system can download the update package in advance and trigger the installation at the precise time point, avoiding conflicts with user operations.

[0055] Through the above technical solutions, the present application effectively solves the problem of high blindness in traditional firmware update timing selection. By combining the probability threshold and time length double screening mechanism, the update operation can neither interfere with the normal use of the user nor complete the installation process. At the same time, the control method to the time point avoids the waste of resources caused by the large time period range in the traditional scheme, such as selecting a 2-hour window, the system can automatically select the precise time point with the lowest network load in the window to start the update, further optimizing the update efficiency.

[0056] The application further proposes an update timing of the firmware update based on the selective confirmation, detects the arrival of the corresponding timing, and controls the automatic execution of the firmware update, including: controlling the pre-download of the firmware update package in the spatial time period and verifying the integrity of the update package; based on the update timing of the firmware update based on the selective confirmation, the specific time point of the update timing is parsed; when the specific time point of the update timing is detected, the downloaded firmware update package is controlled to automatically execute the firmware update; and when the firmware update package update installation is completed, the intelligent terminal is controlled to restart.

[0057] Among them, the control of pre-downloading the firmware update package in the spatial time period means that the download of the update package is completed in advance according to the predicted idle time period of the intelligent terminal, which can be specifically implemented by using a background thread scheduling or a timing task triggered download process to avoid occupying network resources when the user uses the device. The verification of the integrity of the update package means that the file is confirmed to be undamaged or tampered by checking the hash value or digital signature, which can be specifically implemented by using MD5 check or SHA-256 algorithm to ensure the safety of the update process. The parsing of the specific time point of the update timing means that the predicted idle time period is converted into an executable time trigger condition, which can be specifically implemented by using a timestamp conversion or a calendar event parsing module to enable the system to accurately identify the update start time. The control of automatically executing the firmware update means that the installation program is triggered by script calling or system service, which can be specifically implemented by using Shell script or Windows service to realize automatic operation and reduce manual intervention. The control of restarting the intelligent terminal means that the restart instruction is sent after the update is completed, which can be specifically implemented by using operating system API calling or hardware reset signal to ensure that the new firmware takes effect.

[0058] Specifically, in the predicted idle time period, the system downloads the firmware update package through the background process, and immediately performs integrity verification after the download is completed. If the verification fails, the download request is reinitiated until success. When the parsed specific time point is reached, the system calls the preset installation program to perform the update operation, and monitors the process state in real time during the installation process. After the installation is completed, the system automatically triggers the restart instruction to make the device load the new version firmware. For example, in the intelligent television scenario, the system may detect that the intelligent terminal is in an idle state at 2 a.m., at which time the download, verification and installation are automatically completed, and the device is restarted after the update is completed, so that the user can obtain the latest function when using the device the next day.

[0059] Compared with the prior art, the traditional firmware update scheme usually forcibly downloads and installs in a fixed period, which is easy to suddenly pop up an update prompt when the user is watching a video, resulting in interruption of operation. The present scheme avoids interference with the use experience by pre-downloading and precise triggering mechanism, and only performs the update when the device is idle, for example, during the user's sleep period. In addition, the prior art lacks an integrity verification link, which may cause the update package to be damaged due to network fluctuations, and the present scheme effectively prevents this problem through hash verification.

[0060] By the technical solution, the application solves the problem of installation failure caused by unstable network in the firmware updating process, and guarantees the reliability of the update file through integrity verification. At the same time, the accurate update time point control avoids interruption of user operation, and the restart mechanism ensures that the new firmware can take effect in time, so as to realize the automatic updating process with zero interference in the unattended scene.

[0061] The application further proposes that before the step of detecting the arrival of the corresponding time and controlling the automatic execution of the firmware update in the update time of the firmware update based on the selective acknowledgement, the steps of pre-setting the support of breakpoint resume in the firmware updating process are included for avoiding the update failure caused by network interruption; pre-setting the support of rollback mechanism in the firmware updating process for ensuring the recovery to the original version after the update failure; and pre-setting the real-time saving of log records in the firmware updating process for recording the update process and results.

[0062] The breakpoint resume refers to that in the firmware update package transmission process, if the network connection is interrupted, the remaining data can be continuously transmitted from the interrupted position after the recovery of the connection, and the Range header field based on the HTTP protocol or the block transmission encoding mechanism can be specifically used to realize the feature, which can avoid the repeated download or update failure caused by network fluctuation. The rollback mechanism refers to that when the abnormality or verification failure is detected in the firmware update installation process, the system is automatically restored to the stable version before the update, and the backup partition image or version control file can be specifically used to realize the feature, which ensures that the device can still be normally used when the update fails. The log record refers to that the download progress, installation state, error code and timestamp information are recorded in real time in the update process, and the database storage or local file system writing mode can be specifically used to realize the feature, which provides traceable operation records for fault troubleshooting.

[0063] Specifically, when the firmware update is performed in the predicted idle time period, the update package is downloaded in stages through the breakpoint resume mechanism, and if the network is interrupted during the downloading process, the amount of transmitted data is recorded and the remaining part is continuously transmitted after the network is recovered. In the installation stage, if the file verification error or installation timeout is detected, the rollback mechanism is triggered to restore the system to the backup partition before the update. At the same time, each operation step in the update process is recorded in real time to the log file, including the download start time, the breakpoint resume times, the installation result and the rollback event. For example, when the intelligent terminal enters the predicted idle time period at 2 a.m., the system automatically starts the update process, and if the network is unstable at this time, the breakpoint resume function can avoid the re-download of the complete package, thereby saving the bandwidth and time.

[0064] Compared with the prior art, the existing firmware updating scheme usually needs to download the complete update package again in the event of network interruption, which not only wastes resources but also may cause the user equipment to be unable to upgrade in time due to repeated failures. In addition, the traditional method lacks an effective version rollback mechanism, which may cause the equipment to be unable to function normally once the update fails. The present scheme reduces network dependence through breakpoint resuming, guarantees system stability through a rollback mechanism, and provides complete operation tracing capability through log recording, forming multiple fault tolerance safeguards.

[0065] Through the technical scheme described above, the present application can effectively deal with network interruption, installation abnormalities and other sudden problems during the firmware updating process, avoiding termination of the entire updating process due to a single failure. Breakpoint resuming reduces repeated data transmission, the rollback mechanism prevents system crashes, and log recording provides precise fault location basis for operation and maintenance personnel, and the three work together to significantly improve the success rate and reliability of firmware updating.

[0066] The present application will be described in further detail below through another specific application embodiment.

[0067] Embodiment Two The AI-driven firmware updating processing method provided in this embodiment two comprises: S10, analyze user usage habits by obtaining log data and sensor data of the user's intelligent terminal. The log data includes: viewing time, usage frequency, device operation type. The sensor data includes: whether the user is in front of the device detected by the camera.

[0068] This step inputs log data and sensor data, and specifically can use a clustering algorithm (such as K-means) to analyze user behavior patterns. Statistical methods are used to calculate user usage habits (such as average viewing time, peak period). Output the user behavior feature vector as the user usage habit data.

[0069] S11, predict the future idle time period based on user behavior according to the user usage habit data.

[0070] In this embodiment, a time series prediction model (such as LSTM, Prophet) can be used; the user historical usage data, date type (weekday / weekend), holiday information of the user usage habit data are analyzed to analyze the probability distribution of the future idle time period. The optimal model parameters can be selected through cross-validation, and then the probability distribution of the future idle time period is output.

[0071] S12, select the best updating time based on the prediction result of predicting the future idle time period.

[0072] Specifically, the step can preferentially select a period with high idle probability and long duration; and avoid selecting a period in which the user can suddenly use the device. The specific constraint condition can be set as: the update time is not more than the maximum allowed duration; and the update frequency meets the requirements of the device manufacturer.

[0073] That is, in the step, the probability distribution of the future idle period is input, and the optimal update opportunity is output.

[0074] S13, automatically performing firmware update at the selected opportunity.

[0075] Specifically, the application can download the firmware update package in the idle time of the intelligent terminal in advance, and verify the integrity of the update package (such as MD5 verification). Then, the firmware update is automatically performed at the selected opportunity, and the device is restarted after the firmware update is installed.

[0076] In specific implementation, the application can also set a security mechanism: for example, support for breakpoint resume function is set, so as to avoid network interruption leading to update failure. And support for rollback mechanism function is set to ensure that the original version can be restored after the update fails. And at the same time, the update process and result are recorded through log recording, which is convenient for subsequent analysis. The step inputs the optimal update opportunity, and outputs the firmware update result (such as success / failure).

[0077] S14, optimizing the prediction model through user feedback and update result.

[0078] In the embodiment of the step, the AI prediction model is optimized by receiving user feedback information. The user feedback information includes: whether the update is perceived, whether the use is affected; and the update result includes: whether it is successful, whether it causes device abnormality.

[0079] The specific optimization method in the embodiment can include: for example, using incremental learning (Incremental Learning) to update the prediction model; and the priority mechanism can be optimized by introducing reinforcement learning (Reinforcement Learning).

[0080] In the specific embodiment, regarding data security, federated learning (Federated Learning) can be used to protect user privacy, and only the model parameters are updated locally.

[0081] The step inputs the user feedback and update result, and outputs the optimized prediction model.

[0082] The specific application embodiment is specifically shown as follows: example 1, the user behavior analysis module analyzes the viewing habits of the user in the past week, and it is found that the user usually watches TV from 8 pm to 10 pm. The idle time prediction module predicts that the user will be in an idle state from 1 am to 3 am in the next few days. The update timing optimization module selects 1 am as the best update timing. Then the update execution module automatically downloads and installs the firmware update at 1 am. And the feedback optimization module records the update result (such as whether it is successful, whether it affects the user) and optimizes the prediction model.

[0083] Example 2, for example, the user behavior analysis module analyzes the user device operation habits, and it is found that the user uses less frequently on weekends; then the idle time prediction module predicts that the user will be in an idle state from 10 am to 12 pm on weekends. Then the update timing optimization module selects 10 am on Saturday as the best update timing.

[0084] Then the update execution module automatically performs the firmware update at 10 am on Saturday. And the feedback optimization module optimizes the prediction model according to the user feedback (such as whether the update is perceived).

[0085] As can be seen from the above, the present application can achieve the following effects through specific application embodiments: 1) precise prediction can be achieved: because the AI analyzes the user usage habits, the best update timing is accurately predicted. 2) minimum interference can be achieved: because the update is performed during the user's idle period, the use experience is not affected. 3) continuous optimization can be achieved: the prediction model is continuously improved by the feedback optimization module, enhancing the system adaptability. 4) safety is improved: because the present application supports breakpoint resume and rollback mechanism, ensuring the safety and reliability of the update process.

[0086] Exemplary device As Figure 2 shown, the embodiment of the present application provides an AI-driven firmware update processing device, which comprises: A user behavior analysis module 310 is configured to acquire log data and sensor data of the intelligent terminal, and analyze user usage habit data based on the log data and sensor data; An idle time prediction module 320 is configured to predict future idle time periods of the corresponding intelligent terminal based on the analyzed user usage habit data by using a time series AI prediction model; An update timing optimization module 330 is configured to select an update timing for confirming firmware update based on the predicted future idle time periods of the corresponding intelligent terminal; An update execution module 340 is configured to detect the arrival of the corresponding timing based on the selected update timing for confirming firmware update, and control the automatic execution of firmware update; The feedback optimization module 350 is configured to receive feedback information of the user on the firmware update, and optimize the AI prediction model through the feedback information and the firmware update result.

[0087] The user behavior analysis module refers to a composite data collection system composed of a log data analysis unit and a sensor data collection unit. The user behavior analysis module can be implemented by using a distributed data collection framework. The log data analysis unit is responsible for extracting structured data such as viewing time and device operation type. The sensor data collection unit acquires user presence state data by calling a camera API. The idle time prediction module refers to a time series prediction engine constructed based on an LSTM neural network. The idle time prediction module can be implemented by using a TensorFlow framework to train a model. After the input layer receives a user behavior feature vector, the idle time prediction module performs pattern recognition through a bidirectional recurrent neural network layer. The output layer generates a future time period idle probability distribution. The update timing optimization module refers to a decision unit with a probability threshold judgment function. The update timing optimization module can be implemented by using a sliding window algorithm to filter time periods that meet a preset idle probability threshold and have a duration longer than a firmware update required duration. The update execution module refers to an execution chain including a firmware downloader, an integrity checker, and an automatic installer. The firmware downloader uses a fragmentation transmission technology to realize breakpoint resume transmission. The integrity checker verifies the integrity of the update package by comparing hash values. The automatic installer triggers a silent installation process when a preset time point is detected. The feedback optimization module refers to a model optimization component with online learning capability. The feedback optimization module can be implemented by using an incremental learning algorithm to fuse user feedback data and update result data to generate a new training data set, and then perform parameter tuning on the AI prediction model.

[0088] Specifically, the user behavior analysis module continuously acquires device operation records through a log collection interface, and simultaneously calls a camera sensor interface to detect the user presence state. The collected raw data is input into a clustering analysis unit after data cleaning, and a K-means algorithm is used to identify user behavior patterns and calculate the daily average usage time distribution. The idle time prediction module inputs the processed user behavior feature vector into the trained LSTM model in time sequence, and the model outputs the idle probability value of each time period within the next 24 hours. The cross-validation algorithm is used to select the model parameter combination with the smallest prediction error. The update timing optimization module performs sliding window scanning on the prediction results. When it is detected that the idle probability of three consecutive time units all exceeds 85% and the total time length meets the firmware update requirement, the time period is marked as a candidate update window. The update execution module starts the firmware package download task before the candidate window starts, and performs integrity check by calling the SHA-256 algorithm after the download is completed. After the check is passed, it enters the waiting trigger state, and automatically executes the installation program and completes the device restart when the system clock reaches the predetermined time point. The feedback optimization module collects abnormal event records about the update process in the user operation log after the update is completed, generates an optimization data set combining the update time consumption, success rate and other indicators, and dynamically adjusts the weight parameters of the LSTM model through the gradient descent algorithm.

[0089] Compared with the prior art, the traditional firmware update device only relies on a simple timer or a manual trigger mechanism, and cannot effectively identify the real use state of the user, which is easy to mis-trigger the update process during the peak period of device use. The device constructs a user behavior portrait through multi-modal data fusion analysis, accurately locates the low-interference period by using a prediction model with time sequence modeling capability, and forms a closed-loop optimization system combined with a feedback mechanism. For example, the update strategy of the prior art uses a fixed early morning period, which will still cause use interruption when the user stays up to watch, while the device can dynamically adjust the update time according to the actual behavior pattern of the individual user, avoiding the limitations of a single time strategy.

[0090] Through the above technical solutions, the application effectively solves the technical problem that the firmware update process causes interference to the user use, and realizes intelligent selection of the update timing. The device predicts the optimal update time window by real-time analysis of user behavior characteristics, ensures the smooth completion of the update task while minimizing the impact on normal use of the device, continuously optimizes the prediction accuracy with the feedback mechanism, and forms a firmware update management system with adaptive capability.

[0091] Based on the above embodiment, the application further provides an intelligent terminal, and a principle block diagram thereof can be as shown in Figure 3 The intelligent terminal includes a processor, a memory, a network interface, a display screen, and a database connected through a system bus.

[0092] The memory stores one or more programs configured to be executed by the processor to implement a processing method for automatically selecting a firmware update timing based on user usage habits.

[0093] The smart terminal refers to a smart device with data processing capability, such as a smart TV, a smart speaker, or a smart home control terminal. It collects user behavior data through built-in sensors and log recording functions. The memory can be a flash memory, a hard disk, or a cloud storage space, used to store program codes and user behavior data. The processor can be a central processing unit or a special AI chip, used to execute algorithm logic in the program, such as a time series prediction model. The program contains a firmware update method, which includes analyzing user usage habits, predicting idle time periods, selecting update timing, and automatically performing updates. By combining user behavior data with prediction models, the update process is ensured to avoid user active periods.

[0094] Specifically, the smart terminal continuously collects log data and sensor data during operation, such as viewing duration, device operation type, and user presence state detected by the camera. These data are converted into user behavior feature vectors, such as peak hours or average use duration, through clustering algorithms and statistical methods. Then, a time series prediction model generates a probability distribution of future idle time periods based on the feature vectors, and selects a period that meets the idle probability threshold and has sufficient duration as the update timing. When the predetermined time point is reached, the firmware update package is downloaded and verified in the background, and the installation and restart operations are automatically performed after confirming that the user is not using the device.

[0095] In some embodiments, the program can further include a resume function, such as saving the progress of the downloaded update package when the network is interrupted; a rollback mechanism can automatically recover to the original version after the update fails, such as version rollback through backup partition; a log recording function can record key events in the update process, such as download completion time or installation error code, for subsequent analysis and optimization.

[0096] Compared with the prior art, the existing scheme relies on fixed time or manual trigger for update, which is easy to cause interruption when the user uses the device, while the present scheme dynamically adjusts the update timing by analyzing user behavior patterns, such as performing updates during user sleep periods or during weekdays, thereby reducing the interference with user experience. In addition, the prior art lacks comprehensive utilization of sensor data, such as not using the camera to detect whether the user is present, while the present scheme improves prediction accuracy through multi-dimensional data fusion.

[0097] By the technical solution, the application can significantly reduce the interference of firmware update on user use of the device, for example, avoid forcibly restarting the device when the user watches TV, and improve the update success rate. By optimizing the update time selection, reducing the update failure caused by user active cancellation or device occupation, and supporting background silent download and installation, the device performance and security can be upgraded in time.

[0098] The application further proposes a computer readable storage medium, when the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can execute the AI-driven firmware update processing method, which includes obtaining log data and sensor data of the intelligent terminal to analyze user usage habits, using a time series AI prediction model to predict future idle time period, selecting firmware update time according to the prediction result and automatically performing update.

[0099] Among them, the computer readable storage medium refers to a physical carrier that stores computer executable instructions, which can be implemented by solid state disk, flash disk or optical disk, and its function is to convert the logic of the firmware update processing method into program code that can be read and executed by the electronic device. The processor of the electronic device refers to the operation unit that executes instructions, which can be implemented by a central processing unit or a graphics processing unit, and its function is to analyze the instructions in the storage medium and control the automatic operation of the firmware update process.

[0100] Specifically, when the instructions in the storage medium are loaded into the electronic device, the device collects user behavior data through the sensor and analyzes the usage habits, for example, detects whether the user is in front of the device through the camera. The time series AI model predicts the future idle period based on historical data, for example, using the usage mode of weekdays and holidays as input features. The update time is set to the period when the probability threshold exceeds the preset value and the duration meets the requirements, for example, selecting 1am to 3am as the update window. The update package starts installation in the idle period, and ensures the reliability of the process through breakpoint continuation and log recording, for example, automatically saving the progress when the network is interrupted and continuing the update after recovery.

[0101] Compared with the prior art, the traditional scheme relies on fixed time or manual trigger update, which is easy to cause interference in the user active period, while the present scheme can adaptively adjust the update strategy by dynamically analyzing user behavior data and predicting idle time. The prior art lacks comprehensive utilization of sensor data, while the present scheme combines camera detection and log analysis to improve the accuracy of idle time judgment.

[0102] By the technical scheme, the application solves the problem of inaccurate selection of firmware update time, which leads to a decline in user experience, and realizes dynamic matching of the update process and user usage habits. The device only performs updates during the predicted idle period, avoiding accidental interruptions when watching videos or operating the device, while reducing the need for manual intervention through an automated process, improving update efficiency and system stability.

[0103] The above merely illustrates the embodiments of the application and is not used to limit the protection scope of the application. For those skilled in the art, the application can have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the application shall be included in the protection scope of the application.

Claims

1. An AI-driven firmware update processing method, characterized by, The application relates to a method for automatically updating firmware of a smart terminal. The method comprises the following steps: acquiring log data and sensor data of the smart terminal, and analyzing user usage habit data through the log data and the sensor data; adopting an AI prediction model of time sequence to predict a future idle time period of the corresponding smart terminal according to the analyzed user usage habit data; selecting and confirming an updating time of firmware update according to the predicted future idle time period of the corresponding smart terminal; 2.The AI-driven firmware update processing method of claim 1, wherein, detecting that the corresponding time arrives based on the selected and confirmed updating time of firmware update, and controlling automatic execution of firmware update. The method further comprises the following steps after the step of detecting that the corresponding time arrives based on the selected and confirmed updating time of firmware update, and controlling automatic execution of firmware update: 3.The AI-driven firmware update processing method of claim 1, wherein, receiving feedback information of the user on the firmware update, and optimizing the AI prediction model through the feedback information and a firmware update result. The step of acquiring log data and sensor data of the smart terminal, and analyzing user usage habit data through the log data and the sensor data comprises the following steps: acquiring log data and sensor data of the smart terminal, wherein the log data comprises viewing time length data, usage frequency data and device operation type data; and the sensor data comprises data detected by a camera on whether a user is in front of the device; analyzing user behavior patterns by using a clustering algorithm on the log data and the sensor data, and calculating user usage habits by using a statistical method, wherein the usage habits comprise average viewing time length and peak time period; 4.The AI-driven firmware update processing method of claim 1, wherein, outputting a user behavior feature vector as user usage habit data according to the analyzed user behavior patterns and usage habits. The step of adopting an AI prediction model of time sequence to predict a future idle time period of the corresponding smart terminal according to the analyzed user usage habit data comprises the following steps: inputting the user behavior feature vector as user usage habit data into the AI prediction model of time sequence, and outputting a probability distribution of the future idle time period; wherein the user behavior feature vector comprises user historical usage data, date types of weekdays and weekends and legal holiday information; 5.The AI-driven firmware update processing method of claim 1, wherein, obtaining the predicted future idle time period of the corresponding smart terminal by selecting and optimizing model parameters through cross-validation according to the output probability distribution of the future idle time period. The step of selecting and confirming an updating time of firmware update according to the predicted future idle time period of the corresponding smart terminal comprises the following steps: acquiring the predicted future idle time period of the corresponding smart terminal, and selecting a time period with an idle probability reaching a predetermined value and a long duration as the updating time of firmware update; 6.The AI-driven firmware update processing method of claim 1, wherein, wherein the updating time of firmware update is an updating time including a specific time point. The step of detecting that the corresponding time arrives based on the selected and confirmed updating time of firmware update, and controlling automatic execution of firmware update comprises the following steps: controlling to download a firmware update package in a space time period in advance, and verifying the integrity of the update package; analyzing the specific time point of the updating time based on the selected and confirmed updating time of firmware update; controlling to automatically execute firmware update by using the downloaded firmware update package when the specific time point of the updating time is detected; controlling to restart the smart terminal when firmware update package updating and installation are completed. 7.The AI-driven firmware update processing method of claim 6, wherein, The update timing of the firmware update based on the selection confirmation includes the following steps before the step of controlling automatic execution of the firmware update: The firmware update process is set to support breakpoint resume in advance, which is used to avoid update failure caused by network interruption; The firmware update process is set to support rollback mechanism in advance, which is used to ensure that the original version can be restored after the update fails; The firmware update process is set to save log records at any time in advance, which is used to record the update process and results.

8. An AI-driven firmware update processing apparatus, characterized by comprising: The device comprises: A user behavior analysis module is configured to obtain log data and sensor data of the intelligent terminal, and analyze user usage habit data based on the log data and sensor data; An idle time prediction module is configured to predict future idle time periods of the corresponding intelligent terminal based on the analyzed user usage habit data by using an AI prediction model of time sequence; An update timing optimization module is configured to select the update timing of the firmware update based on the predicted future idle time periods of the corresponding intelligent terminal; An update execution module is configured to detect the arrival of the corresponding timing based on the selected update timing of the firmware update based on the selection confirmation, and control automatic execution of the firmware update; A feedback optimization module is configured to receive feedback information of the firmware update from the user, and optimize the AI prediction model based on the feedback information and the firmware update results.

9. A user terminal, characterized in that A memory and one or more programs are included, wherein the one or more programs are stored in the memory and configured to be executed by one or more processors, and the one or more programs include a method as claimed in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, When the instructions in the storage medium are executed by the processor of the electronic device, the electronic device can perform the method as claimed in any one of claims 1-7.

Images