Intelligent power saving method and device based on user behavior prediction, equipment, medium and program product

Abstract

The invention relates to the technical field of battery management, in particular to an intelligent power saving method and device based on user behavior prediction, equipment, a medium and a program product. The method comprises the following steps: in response to interaction data detected by target equipment, processing the interaction data based on a pre-constructed behavior prediction model to obtain behavior prediction information of the target equipment; based on the behavior prediction information, determining target hardware units associated with a target application which is started at a high probability, and a hardware active sequence of each target hardware unit; determining a plurality of aligned windows based on the plurality of hardware active sequences; and replacing the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, performing centralized wake-up on the target hardware unit in the alignment window, and maintaining a low-power-consumption state for the target hardware unit outside the alignment window. By adopting the method, the adaptability of power management can be improved, and the endurance performance is improved.

Description

Technical Field

[0001] This application relates to the field of battery management technology, and in particular to an intelligent power-saving method, apparatus, device, medium, and program product based on user behavior prediction. Background Technology

[0002] With the widespread use of portable electronic devices, lithium batteries, with their advantages of high energy density, long cycle life, low self-discharge rate, and no memory effect, have become the core energy storage component of power banks (or portable chargers). From smartphones and tablets to Bluetooth headsets and wearable devices, power banks, as indispensable portable energy replenishment stations, rely directly on the effective management of their internal lithium battery packs for performance and safety. However, the inherent chemical characteristics of lithium batteries also impose strict requirements on their application. For example, their operating voltage must be stable within a specific range; overcharging, over-discharging, overcurrent, or high-temperature operation can all cause permanent degradation of battery performance and even lead to safety hazards such as thermal runaway. Therefore, an efficient and reliable Battery Management System (BMS) is crucial for ensuring the overall performance, lifespan, and user safety of power banks.

[0003] In the current application landscape, the portable power bank market is characterized by diversification and widespread adoption. Product capacities range from several thousand mAh to tens of thousands of mAh to meet the battery life needs of various devices. Simultaneously, the rise of fast charging technology has placed higher demands on the charging and discharging efficiency and management precision of portable power banks. However, intense market competition and significant cost control pressures have resulted in some mid-to-low-end products employing relatively basic battery management solutions, leaving room for improvement in terms of accuracy, functionality, and safety. In daily use, users often encounter problems such as inaccurate power level displays, slow charging speeds, significant capacity degradation after prolonged use, or overheating under extreme conditions. These issues are all closely related to the quality of battery management technology.

[0004] In related technologies, basic lithium battery power bank management solutions, through integrated management ICs and peripheral circuits, achieve necessary charging and discharging control and safety protection, meeting the basic usage requirements of products. However, these solutions still have limitations in terms of power metering accuracy, adaptive management under complex operating conditions, multi-cell balancing control, and thermal management efficiency, which restricts further improvement in the performance and safety of power bank products and needs to be optimized. Summary of the Invention

[0005] Therefore, it is necessary to provide an intelligent power-saving method, device, computer equipment, computer-readable storage medium, and computer program product based on user behavior prediction that can improve the adaptability of power management and improve battery life, thereby addressing the aforementioned technical problems.

[0006] Firstly, this application provides an intelligent power-saving method based on user behavior prediction. The method includes: In response to the interaction data detected by the target device, the interaction data is processed based on a pre-built behavior prediction model to obtain the behavior prediction information of the target device. The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device. Based on the behavioral prediction information, target hardware units associated with target applications that have a high probability of being launched, and hardware activity sequences of each target hardware unit are determined. Several alignment windows are determined based on several hardware active sequences, each alignment window including a wake-up time point, and the alignment window is associated with the target hardware unit; The wake-up time of the target hardware unit is replaced with the wake-up time point of the associated alignment window. The target hardware unit is centrally woken up within the alignment window, and the target hardware unit is kept in a low-power state outside the alignment window.

[0007] In one embodiment, determining a plurality of alignment windows based on a plurality of hardware active sequences, wherein the alignment windows include wake-up time points, and the alignment windows are associated with the target hardware unit, including: Determine the time window point corresponding to each hardware access request in the hardware active sequence, and perform clustering processing using the distance between two time window points and the minimum number of requests required to form a cluster as clustering parameters to obtain the clustering result; The alignment window is generated based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task. In one embodiment, the alignment window is generated based on the clustering result, and the alignment window includes window open time, window close time, window length, and window task: Based on the clustering results, the maximum value of the time window in each cluster is determined, and the maximum value of the time window is set as the baseline window length. A preset start time is added to the beginning of the reference window length, and a preset end time is added to the end of the reference window length to obtain the window length. In one embodiment, replacing the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, performing centralized wake-up of the target hardware unit within the alignment window, and maintaining a low-power state for the target hardware unit outside the alignment window includes: Based on the behavior prediction information, the time sensitivity of the target application is determined, and real-time tasks with time sensitivity higher than a preset threshold are selected. The real-time task is processed independently, and the target hardware unit is woken up at a precise time point corresponding to the real-time task. In one embodiment, the method further includes: Acquire the power consumption log data and historical interaction data of the target device within a preset historical period. The power consumption log data includes power monitoring data and associated hardware status and application behavior sequences. A training set is constructed based on the electricity consumption log data and the historical interaction data to build an initial machine learning model. The initial machine learning model is then trained to convergence using the training set to obtain the behavior prediction model. In one embodiment, the step of constructing a training set based on the electricity consumption log data and the historical interaction data, constructing an initial machine learning model, and training the initial machine learning model with the training set until convergence to obtain the behavior prediction model includes: Several different prediction time windows are set, and several independent behavior prediction models are trained based on the different prediction time windows; A confidence level assessment is performed on each behavior prediction model to obtain a prediction performance index for each behavior prediction model. The prediction time window corresponding to the maximum value of the prediction performance index is selected to obtain the prediction result of the target device.

[0008] Secondly, this application also provides an intelligent power-saving device based on user behavior prediction. The device includes: The behavior prediction module is used to respond to the interaction data detected by the target device, process the interaction data based on the pre-built behavior prediction model, and obtain the behavior prediction information of the target device. The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device. The information parsing module is used to determine, based on the behavior prediction information, the target hardware units associated with the target application that has a high probability of being launched, and the hardware activity sequence of each target hardware unit. An alignment window module is used to determine a plurality of alignment windows based on a plurality of hardware active sequences, wherein the alignment window includes a wake-up time point and the alignment window is associated with the target hardware unit; The power control module is used to replace the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, to centrally wake up the target hardware unit within the alignment window, and to maintain the target hardware unit in a low-power state outside the alignment window.

[0009] In one embodiment, the alignment window module includes: The clustering analysis module is used to determine the time window point corresponding to each hardware access request in the hardware active sequence, and to perform clustering processing using the distance between two time window points and the minimum number of requests required to form a cluster as clustering parameters to obtain the clustering result. The window generation module is used to generate the alignment window based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task. In one embodiment, the window generation module includes: A baseline window module is used to determine the maximum value of the time window in each cluster based on the clustering results, and set the maximum value of the time window as the baseline window length. The window length determination module is used to add a preset start time to the beginning of the reference window length and a preset end time to the end of the reference window length to obtain the window length. In one embodiment, the power control module includes: The real-time task module is used to determine the time sensitivity of the target application based on the behavior prediction information and filter out real-time tasks whose time sensitivity is higher than a preset threshold. An independent processing module is used to process the real-time task independently and wake up the target hardware unit at a precise time point corresponding to the real-time task. In one embodiment, the device further includes: The historical data module is used to acquire the power consumption log data and historical interaction data of the target device within a preset historical period. The power consumption log data includes power monitoring data and related hardware status and application behavior sequences. The model training module is used to construct a training set based on the electricity consumption log data and the historical interaction data, construct an initial machine learning model, and train the initial machine learning model with the training set until convergence to obtain the behavior prediction model. In one embodiment, the model training module includes: A multi-timescale module is used to set several different prediction time windows, and to train several independent behavior prediction models based on different prediction time windows; The prediction window evaluation module is used to evaluate the confidence level of each behavior prediction model, obtain the prediction performance index of each behavior prediction model, and select the prediction time window corresponding to the maximum value of the prediction performance index to obtain the prediction result of the target device.

[0010] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of an intelligent power-saving method based on user behavior prediction as described in any embodiment of the first aspect.

[0011] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of an intelligent power-saving method based on user behavior prediction as described in any embodiment of the first aspect.

[0012] Fifthly, this application also provides a computer program product. The computer program product includes a computer program that, when executed by a processor, implements the steps of an intelligent power-saving method based on user behavior prediction as described in any embodiment of the first aspect.

[0013] The aforementioned intelligent power-saving method, device, computer equipment, storage medium, and computer program product based on user behavior prediction, derived through technical features, can achieve beneficial effects to address the technical problems in the background art: This application provides an intelligent power-saving method based on user behavior prediction. The method includes: responding to interaction data detected by a target device; processing the interaction data based on a pre-built behavior prediction model to obtain behavior prediction information of the target device, wherein the behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device; determining target hardware units associated with target applications with a high probability of being launched, and hardware activity sequences of each target hardware unit, based on the behavior prediction information; determining several alignment windows based on several hardware activity sequences, wherein the alignment window includes a wake-up time point, and the alignment window is associated with the target hardware unit; replacing the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window; centrally waking up the target hardware unit within the alignment window; and maintaining the target hardware unit in a low-power state outside the alignment window. In implementation, the pre-built behavior prediction model enables real-time analysis of device interaction data, which helps to predict the user's next application usage intention. Architecturally, this changes the passive response mode of traditional power-saving technologies, elevating power management from "post-event remediation" to "pre-event planning," thereby helping to lay a decision-making foundation for subsequent power resource regulation. Subsequently, the system maps abstract behavioral prediction information to specific target hardware units and their active time sequences. This helps to directly link user behavior with the underlying hardware power consumption, enabling power-saving strategies to extend from the application layer to the hardware layer. Furthermore, based on these dispersed hardware active sequences, several optimized "alignment windows" are dynamically calculated. This facilitates the intelligent merging and alignment of previously scattered and independent hardware wake-up points through clustering and scheduling algorithms. Ultimately, on the one hand, the wake-up time of each alignment window is as close as possible to the actual hardware requirements, helping to reduce performance latency or power waste caused by premature or high-frequency wake-ups. On the other hand, by dynamically adjusting the window length, it helps to precisely meet the needs of the tasks within the window, reducing the possibility of idle losses due to insufficient resource allocation or excessively long windows. This allows the device to transition from a high-power state of "frequent brief wake-ups" to an intelligent control state of "long-term deep sleep and occasional efficient concentrated work." Since the power consumption of electronic devices during state switching (especially from sleep to active) is much higher than the steady-state power consumption, significantly reducing the number of wake-ups can bring substantial power-saving benefits and improve the overall battery life of the device. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 This is an application environment diagram of an intelligent power-saving method based on user behavior prediction in one embodiment; Figure 2 This is a schematic diagram of the first process of an intelligent power-saving method based on user behavior prediction in one embodiment; Figure 3 This is a schematic diagram of the second process of a smart power-saving method based on user behavior prediction in another embodiment; Figure 4 This is a schematic diagram of the third process of an intelligent power-saving method based on user behavior prediction in another embodiment; Figure 5 This is a schematic diagram of the fourth process of an intelligent power-saving method based on user behavior prediction in another embodiment; Figure 6 This is a schematic diagram of the fifth process of an intelligent power-saving method based on user behavior prediction in another embodiment; Figure 7 This is a schematic diagram of the sixth process of an intelligent power-saving method based on user behavior prediction in another embodiment; Figure 8 This is a structural block diagram of an intelligent power-saving device based on user behavior prediction in one embodiment; Figure 9 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0017] This application provides an intelligent power-saving method based on user behavior prediction, which can be applied to, for example... Figure 1In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or located on the cloud or other network servers. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, projection devices, etc. Portable wearable devices can include smartwatches, smart bracelets, head-mounted devices, etc. Head-mounted devices can be virtual reality (VR) devices, augmented reality (AR) devices, smart glasses, etc. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0018] In one embodiment, such as Figure 2 As shown, a smart power-saving method based on user behavior prediction is provided, which can be applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps: Step 202: In response to the interaction data detected by the target device, process the interaction data based on the pre-built behavior prediction model to obtain the behavior prediction information of the target device.

[0019] The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device.

[0020] For example, the terminal can obtain behavior prediction information through a behavior prediction model. The behavior prediction information can be a sorted list of (application / state, probability of occurrence), such as [(App_Navigation, 0.75), (App_Music, 0.15), (State_Lock Screen, 0.08), (App_SMS, 0.02)]. Specifically, when obtaining behavior prediction information, the terminal can determine an effective prediction time window to avoid invalid results caused by predictions that exceed the time frame.

[0021] Step 204: Based on the behavior prediction information, determine the target hardware units associated with the target application that has a high probability of being launched, and the hardware activity sequence of each target hardware unit.

[0022] For example, a terminal can determine the hardware resources required for different application activities through behavior prediction information, which can be expressed as a hardware resource demand vector, such as: {CPU: "High performance", GPU: "Medium", Screen: "Always on + High brightness", GPS: "High precision", Network: "Continuous connection"}, {CPU: "Low power consumption", GPU: "Low", Screen: "Off", GPS: "Off", Network: "Intermittent"}, etc. Then, arranging multiple consecutive hardware resource demand vectors according to time yields a hardware activity sequence that identifies different hardware needs.

[0023] Step 206: Determine several alignment windows based on several of the aforementioned hardware active sequences.

[0024] The alignment window includes a wake-up time point and is associated with the target hardware unit.

[0025] Step 208: Replace the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, perform centralized wake-up of the target hardware unit within the alignment window, and maintain a low-power state for the target hardware unit outside the alignment window.

[0026] In the aforementioned intelligent power-saving method based on user behavior prediction, by combining the technical features in the embodiments and making reasonable derivations, the beneficial effect of solving the technical problems raised in the background art is achieved: This application provides an intelligent power-saving method based on user behavior prediction. The method includes: responding to interaction data detected by a target device; processing the interaction data based on a pre-built behavior prediction model to obtain behavior prediction information of the target device, wherein the behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device; determining target hardware units associated with target applications with a high probability of being launched, and hardware activity sequences of each target hardware unit, based on the behavior prediction information; determining several alignment windows based on several hardware activity sequences, wherein the alignment window includes a wake-up time point, and the alignment window is associated with the target hardware unit; replacing the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window; centrally waking up the target hardware unit within the alignment window; and maintaining the target hardware unit in a low-power state outside the alignment window. In implementation, the pre-built behavior prediction model enables real-time analysis of device interaction data, which helps to predict the user's next application usage intention. Architecturally, this changes the passive response mode of traditional power-saving technologies, elevating power management from "post-event remediation" to "pre-event planning," thereby helping to lay a decision-making foundation for subsequent power resource regulation. Subsequently, the system maps abstract behavioral prediction information to specific target hardware units and their active time sequences. This helps to directly link user behavior with the underlying hardware power consumption, enabling power-saving strategies to extend from the application layer to the hardware layer. Furthermore, based on these dispersed hardware active sequences, several optimized "alignment windows" are dynamically calculated. This facilitates the intelligent merging and alignment of previously scattered and independent hardware wake-up points through clustering and scheduling algorithms. Ultimately, on the one hand, the wake-up time of each alignment window is as close as possible to the actual hardware requirements, helping to reduce performance latency or power waste caused by premature or high-frequency wake-ups. On the other hand, by dynamically adjusting the window length, it helps to precisely meet the needs of the tasks within the window, reducing the possibility of idle losses due to insufficient resource allocation or excessively long windows. This allows the device to transition from a high-power state of "frequent brief wake-ups" to an intelligent control state of "long-term deep sleep and occasional efficient concentrated work." Since the power consumption of electronic devices during state switching (especially from sleep to active) is much higher than the steady-state power consumption, significantly reducing the number of wake-ups can bring substantial power-saving benefits and improve the overall battery life of the device.

[0027] In one embodiment, such as Figure 3 As shown, step 206 includes: Step 302: Determine the time window point corresponding to each hardware access request in the hardware active sequence, and perform clustering processing using the distance between two time window points and the minimum number of requests required to form a cluster as clustering parameters to obtain the clustering result.

[0028] Step 304: Generate the alignment window based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task.

[0029] In this embodiment, extracting the time window points of each hardware access request helps transform the complex comparison of time series intervals into a more efficient time point clustering problem. Subsequently, using the spacing between the points in the time window and the minimum number of requests required to form a cluster as key clustering parameters helps intelligently identify dense hardware access clusters in the time dimension, thereby adaptively discovering the hidden, natural hardware usage rhythm behind user behavior, rather than relying on preset, fixed time windows. Another clustering parameter, the "minimum number of requests," serves as an adjustable threshold, effectively filtering out occasional, isolated access requests, helping to prevent the generation of inefficient micro-windows and ensuring the scalability of the clustering results. Then, based on the above clustering results, the final alignment window is generated. Instead of simply using the cluster center as the window point, it traces back and integrates the complete time boundaries of the original hardware activity sequence, contributing to the effectiveness of the alignment window.

[0030] In one embodiment, such as Figure 4 As shown, step 304 includes: Step 402: Based on the clustering results, determine the maximum value of the time window in each cluster, and set the maximum value of the time window as the baseline window length.

[0031] Step 404: Add a preset start time to the beginning of the reference window length and a preset end time to the end of the reference window length to obtain the window length.

[0032] In this embodiment, based on the clustering analysis results, the maximum value of the original time window for all hardware access requests in each cluster is established as the baseline window length. This helps reduce the possibility of critical issues such as forced hardware interruption and task execution failure due to insufficient window length. Subsequently, by introducing a buffer mechanism for startup and shutdown time, by adding startup time at the beginning of the baseline window, the system reserves sufficient preparation time for hardware units to switch from sleep state to working state (e.g., CPU frequency increase, initial GPS module positioning), ensuring that the hardware is ready when the window opens, thereby improving the immediacy and smoothness of task execution. At the same time, adding shutdown time at the end provides the necessary operational margin for data writing, state saving, and safe transition of hardware back to low-power state, helping to prevent data loss or hardware damage caused by abrupt power outages.

[0033] In one embodiment, such as Figure 5 As shown, step 208 includes: Step 502: Determine the time sensitivity of the target application based on the behavior prediction information, and filter out real-time tasks whose time sensitivity is higher than a preset threshold.

[0034] Step 504: Process the real-time task independently and wake up the target hardware unit at the precise time point corresponding to the real-time task.

[0035] In this embodiment, isolating time-sensitive real-time tasks during the alignment wake-up process helps maintain the stability of real-time task execution.

[0036] In one embodiment, such as Figure 6 As shown, the method includes: Step 602: Obtain the power consumption log data and historical interaction data of the target device within a preset historical period. The power consumption log data includes power monitoring data and related hardware status and application behavior sequences.

[0037] Step 604: Construct a training set based on the electricity consumption log data and the historical interaction data, build an initial machine learning model, and train the initial machine learning model with the training set until convergence to obtain the behavior prediction model.

[0038] In this embodiment, training the behavior prediction model using historical data helps improve the application effect of the behavior prediction model.

[0039] In one embodiment, such as Figure 7 As shown, step 604 includes: Step 702: Set several different prediction time windows, and train several independent behavior prediction models based on the different prediction time windows.

[0040] Step 704: Calculate the confidence level of each behavior prediction model to obtain the prediction performance index of each behavior prediction model, and select the prediction time window corresponding to the maximum value of the prediction performance index to obtain the prediction result of the target device.

[0041] In this embodiment, by setting multiple different prediction time windows and training independent prediction models for each, it is helpful to evaluate the prediction time windows based on the objective law that the accuracy of user behavior prediction will naturally decrease as the prediction time increases. This helps to determine the longest possible span that can maintain the highest prediction confidence, which helps to improve the overall energy efficiency of the system.

[0042] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0043] Based on the same inventive concept, this application also provides an intelligent power-saving device based on user behavior prediction for implementing the aforementioned intelligent power-saving method based on user behavior prediction. The solution provided by this device is similar to the solution described in the above method. Therefore, the specific limitations of one or more embodiments of an intelligent power-saving device based on user behavior prediction provided below can be found in the limitations of the intelligent power-saving method based on user behavior prediction described above, and will not be repeated here.

[0044] In one embodiment, such as Figure 8 As shown, an intelligent power-saving device based on user behavior prediction is provided, including: a behavior prediction module, an information parsing module, an alignment window module, and a power consumption control module, wherein: The behavior prediction module is used to respond to the interaction data detected by the target device, process the interaction data based on the pre-built behavior prediction model, and obtain the behavior prediction information of the target device. The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device. The information parsing module is used to determine, based on the behavior prediction information, the target hardware units associated with the target application that has a high probability of being launched, and the hardware activity sequence of each target hardware unit. An alignment window module is used to determine a plurality of alignment windows based on a plurality of hardware active sequences, wherein the alignment window includes a wake-up time point and the alignment window is associated with the target hardware unit; The power control module is used to replace the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, to centrally wake up the target hardware unit within the alignment window, and to maintain the target hardware unit in a low-power state outside the alignment window.

[0045] In one embodiment, the alignment window module includes: The clustering analysis module is used to determine the time window point corresponding to each hardware access request in the hardware active sequence, and to perform clustering processing using the distance between two time window points and the minimum number of requests required to form a cluster as clustering parameters to obtain the clustering result. The window generation module is used to generate the alignment window based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task. In one embodiment, the window generation module includes: A baseline window module is used to determine the maximum value of the time window in each cluster based on the clustering results, and set the maximum value of the time window as the baseline window length. The window length determination module is used to add a preset start time to the beginning of the reference window length and a preset end time to the end of the reference window length to obtain the window length. In one embodiment, the power control module includes: The real-time task module is used to determine the time sensitivity of the target application based on the behavior prediction information and filter out real-time tasks whose time sensitivity is higher than a preset threshold. An independent processing module is used to process the real-time task independently and wake up the target hardware unit at a precise time point corresponding to the real-time task. In one embodiment, the device further includes: The historical data module is used to acquire the power consumption log data and historical interaction data of the target device within a preset historical period. The power consumption log data includes power monitoring data and related hardware status and application behavior sequences. The model training module is used to construct a training set based on the electricity consumption log data and the historical interaction data, construct an initial machine learning model, and train the initial machine learning model with the training set until convergence to obtain the behavior prediction model. In one embodiment, the model training module includes: A multi-timescale module is used to set several different prediction time windows, and to train several independent behavior prediction models based on different prediction time windows; The prediction window evaluation module is used to evaluate the confidence level of each behavior prediction model, obtain the prediction performance index of each behavior prediction model, and select the prediction time window corresponding to the maximum value of the prediction performance index to obtain the prediction result of the target device.

[0046] The modules in the aforementioned intelligent power-saving device based on user behavior prediction can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0047] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 9 As shown, the computer device includes a processor, memory, input / output interfaces, a communication interface, a display unit, and an input device. The processor, memory, and input / output interfaces are connected via a system bus, and the communication interface, display unit, and input device are also connected to the system bus via the input / output interfaces. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The input / output interfaces are used for exchanging information between the processor and external devices. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements an intelligent power-saving method based on user behavior prediction. The display unit is used to form a visually visible image and can be a display screen, a projection device, or a virtual reality imaging device. The display screen can be an LCD screen or an e-ink screen. The input device of the computer device can be a touch layer covering the display screen, or buttons, trackballs, or touchpads set on the casing of the computer device, or external keyboards, touchpads, or mice, etc.

[0048] Those skilled in the art will understand that Figure 9 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0049] In one embodiment, a computer device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in the above method embodiments.

[0050] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon that, when executed by a processor, implements the steps in the above method embodiments.

[0051] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the above method embodiments.

[0052] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.

[0053] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0054] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0055] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A smart power-saving method based on user behavior prediction, characterized in that, The method includes: In response to the interaction data detected by the target device, the interaction data is processed based on a pre-built behavior prediction model to obtain the behavior prediction information of the target device. The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device. Based on the behavioral prediction information, target hardware units associated with target applications that have a high probability of being launched are identified, as well as the hardware activity sequence of each target hardware unit. Several alignment windows are determined based on several hardware active sequences, each alignment window including a wake-up time point, and the alignment window is associated with the target hardware unit; The wake-up time of the target hardware unit is replaced with the wake-up time point of the associated alignment window. The target hardware unit is centrally woken up within the alignment window, and the target hardware unit is kept in a low-power state outside the alignment window.

2. The method according to claim 1, characterized in that, The step of determining several alignment windows based on several hardware active sequences, wherein the alignment windows include wake-up time points, and the alignment windows are associated with the target hardware unit, including: Determine the time window point corresponding to each hardware access request in the hardware active sequence, and perform clustering processing using the distance between two time window points and the minimum number of requests required to form a cluster as clustering parameters to obtain the clustering result; The alignment window is generated based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task.

3. The method according to claim 2, characterized in that, The alignment window is generated based on the clustering results. The alignment window includes window opening time, window closing time, window length, and window task. Based on the clustering results, the maximum value of the time window in each cluster is determined, and the maximum value of the time window is set as the baseline window length. A preset start time is added to the beginning of the reference window length, and a preset end time is added to the end of the reference window length to obtain the window length.

4. The method according to claim 1, characterized in that, The step of replacing the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, performing centralized wake-up of the target hardware unit within the alignment window, and maintaining a low-power state for the target hardware unit outside the alignment window includes: Based on the behavior prediction information, the time sensitivity of the target application is determined, and real-time tasks with time sensitivity higher than a preset threshold are selected. The real-time task is processed independently, and the target hardware unit is woken up at a precise time point corresponding to the real-time task.

5. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Acquire the power consumption log data and historical interaction data of the target device within a preset historical period. The power consumption log data includes power monitoring data and associated hardware status and application behavior sequences. A training set is constructed based on the electricity consumption log data and the historical interaction data to build an initial machine learning model. The initial machine learning model is then trained to convergence using the training set to obtain the behavior prediction model.

6. The method according to claim 5, characterized in that, The step of constructing a training set based on the electricity consumption log data and the historical interaction data, building an initial machine learning model, and training the initial machine learning model to convergence using the training set to obtain the behavior prediction model includes: Several different prediction time windows are set, and several independent behavior prediction models are trained based on the different prediction time windows; A confidence level assessment is performed on each behavior prediction model to obtain a prediction performance index for each behavior prediction model. The prediction time window corresponding to the maximum value of the prediction performance index is selected to obtain the prediction result of the target device.

7. A smart power-saving device based on user behavior prediction, characterized in that, The device includes: The behavior prediction module is used to respond to the interaction data detected by the target device, process the interaction data based on the pre-built behavior prediction model, and obtain the behavior prediction information of the target device. The behavior prediction information is the prediction result of the behavior prediction model for the next set of application states of the target device. The information parsing module is used to determine, based on the behavior prediction information, the target hardware units associated with the target application that has a high probability of being launched, and the hardware activity sequence of each target hardware unit. An alignment window module is used to determine a plurality of alignment windows based on a plurality of hardware active sequences, wherein the alignment window includes a wake-up time point and the alignment window is associated with the target hardware unit; The power control module is used to replace the wake-up time of the target hardware unit with the wake-up time point of the associated alignment window, to centrally wake up the target hardware unit within the alignment window, and to maintain the target hardware unit in a low-power state outside the alignment window.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images