Power supply state self-adaptive switching method and system of intelligent terminal equipment

Abstract

The invention provides a power state self-adaptive switching method and system for intelligent terminal equipment, and the method comprises the steps: obtaining the characteristics of different hardware modules of the intelligent terminal equipment, and determining a multi-stage power state switching process of each hardware module, obtaining specific control steps and parameter adjustment ranges of three stages of pre-switching, switching neutralization and switching completion of each hardware module; setting a local power management state adaptive switching strategy; when power supply management state self-adaptive switching is carried out, a mode switching smooth transition algorithm is adopted according to a multi-stage power supply state switching process of each hardware module and specific control steps and parameter adjustment ranges of three stages of pre-switching, switching neutralization and switching completion of each module; stable transition of various parameters is achieved through gradual adjustment, and power consumption fluctuation caused by large-amplitude parameter changes is avoided.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a method and system for adaptively switching the power state of an intelligent terminal device. Background Art

[0002] With the wide application of mobile devices and Internet of Things devices, while providing efficient and convenient services, they also face challenges such as large power consumption, short battery life, and serious energy consumption. In terms of power management technology, existing power management strategies often rely on fixed preset values or simple state determination methods, and cannot flexibly self-adjust according to the dynamic changes in the actual usage scenarios of the devices. Especially when an intelligent terminal device needs to switch its working state under different working modes (such as light load mode, heavy load mode, sleep mode, etc.), the traditional power management mode often cannot adapt to this dynamic change, resulting in low energy utilization efficiency and fluctuations in device performance. Summary of the Invention

[0003] The present invention provides a method and system for adaptively switching the power state of an intelligent terminal device, aiming to solve the problem that existing power management technologies cannot adaptively respond to changes in the external environment and internal hardware states.

[0004] The present invention provides a method for adaptively switching the power state of an intelligent terminal device, including: Obtaining the characteristics of different hardware modules of the intelligent terminal device, determining the multi-stage power state switching process of each hardware module, and obtaining the specific control steps and parameter adjustment ranges for the pre-switching, switching, and post-switching stages of each hardware module; Setting a local power management state adaptive switching strategy; When performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges for the pre-switching, switching, and post-switching stages of each module, using a mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

[0005] Preferably, the obtaining the characteristics of different hardware modules of the intelligent terminal device, determining the multi-stage power state switching process of each hardware module, and obtaining the specific control steps and parameter adjustment ranges for the pre-switching, switching, and post-switching stages of each hardware module includes: Obtaining the characteristic parameters of at least one hardware module, where the characteristic parameters include power consumption and response time; According to the characteristic parameters, determining the power state adaptive switching process of the intelligent terminal device and the duration of each stage; Obtain power consumption data during the power supply switching process, and use a clustering algorithm to analyze the power consumption data to obtain the specific control steps and parameter adjustment ranges of the hardware module in the three stages of pre-switching, in-switching, and post-switching, so that when performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges of the three stages of pre-switching, in-switching, and post-switching of each module, adopt a mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

[0006] Preferably, the setting of the local power management state adaptive switching strategy includes: Collect one or more power energy states of the power supply; Compare with a preset power energy state threshold to determine whether it is lower than the preset power energy state threshold; When the collected power energy state is lower than the preset power energy state threshold, the device enters the low power mode; Process the historical data sampled in the low power mode, and use the processing result as the input variable of the model to train the neural network model, and feedback the prediction result output by the obtained neural network model to the energy control system; The energy control system controls the energy switching module to switch the device to the solar power supply mode or the wind power supply mode or the light power supply mode or the electric power supply mode according to the prediction result.

[0007] Preferably, it further includes: When the collected power energy state is higher than the preset power energy state threshold, enter the high power mode; Process the historical data sampled in the high power mode, and use the processing result as the input variable of the model to train the neural network model, and feedback the prediction result output by the obtained neural network model to the energy control system. The energy control system controls the energy switching module to switch the device from the solar power supply mode or the wind power supply mode or the light power supply mode or the electric power supply mode to the intelligent scheduling mode according to the prediction result.

[0008] Preferably, it further includes: When the collected power energy state is equal to the preset power energy state threshold, enter the intelligent scheduling mode; Among them, in the intelligent scheduling mode, preferentially use the collected solar power generation, wind power generation, light power generation or electric power for power supply.

[0009] Preferably, when performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges of the three stages of pre-switching, in-switching, and post-switching of each module, adopt a mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment, including: When performing adaptive switching of the power management state, in the pre-switching stage, by analyzing the differences between the hardware module state and the target power state, calculate the variation range of the power state parameters that need to be adjusted, and generate a parameter adjustment plan; After entering the in-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, adopt a mode switching smooth transition algorithm, and perform multiple small-scale voltage and frequency adjustments through a progressive parameter adjustment strategy to achieve a smooth transition of the hardware module state; Enter the post-switching completion stage, and by detecting the hardware module state a preset number of times, determine whether the target power state has been reached. If the state is stable, then officially enter the new power state; if the state is still unstable, then return to the in-switching stage to continue with fine parameter tuning.

[0010] Preferably, it further includes: When performing parameter adjustment in the in-switching stage, monitor the power consumption change of the hardware module in real time. By comparing with a preset power consumption threshold, if the actual power consumption exceeds the threshold, then reduce the adjustment speed or the adjustment range.

[0011] Preferably, when performing adaptive switching of the power management state, in the pre-switching stage, by analyzing the differences between the hardware module state and the target power state, calculate the variation range of the power state parameters that need to be adjusted, and generate a parameter adjustment plan, including: When performing adaptive switching of the power management state, in the pre-switching stage, obtain the operating state parameters of the current hardware. The operating state parameters include the current voltage and frequency parameters, and form a current hardware module state vector according to the operating state parameters; Obtain the target voltage and target frequency parameters in the target power state, and form a target power state vector according to the target voltage and target frequency parameters; Adopt a support vector machine algorithm, use the current hardware module state vector as the input and the target power state vector as the output, and train to form a mapping model from the hardware module state to the power state; Input the current hardware module state vector into the mapping model to obtain the adjustment range of the voltage and frequency parameters required to switch to the target power state in the current hardware module state; According to the adjustment range of the voltage and frequency parameters, formulate a parameter adjustment plan. The parameter adjustment plan includes the adjustment time points, adjustment step sizes, and adjustment durations of each parameter; Adopt a decision tree algorithm, use the current values and target values of the voltage and frequency parameters as features, and use whether the parameters can be adjusted in place as the judgment basis. If the adjustment plan is feasible, then initiate pre-switching; if the adjustment plan is not feasible, then adjust the parameter adjustment plan until the adjustment plan is feasible.

[0012] Preferably, after entering the mid-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, a mode switching smooth transition algorithm is adopted, and multiple small-amplitude voltage and frequency adjustments are carried out through a progressive parameter adjustment strategy to achieve a smooth transition of the hardware module state, including: After entering the mid-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, obtain the voltage and frequency parameters to be adjusted, as well as the amplitude and number of adjustments each time; Judge whether the current hardware module state meets the switching condition. If it meets, enter the mid-switching stage; otherwise, continue to monitor the hardware module state until the switching condition is met; In the mid-switching stage, according to the parameters in the parameter adjustment plan, make the first small-amplitude voltage adjustment, and obtain the current hardware power consumption value after the adjustment; Compare the adjusted power consumption value with the power consumption value before the adjustment, calculate the power consumption fluctuation amplitude, and judge whether it exceeds the preset power consumption fluctuation threshold; If the power consumption fluctuation amplitude does not exceed the threshold, make the next small-amplitude voltage adjustment according to the parameter adjustment plan until all voltage adjustments are completed; After all voltage adjustments are completed, make multiple small-amplitude adjustments to the frequency parameters in the same way, and monitor the power consumption fluctuation amplitude after each adjustment; When all progressive adjustments of voltage and frequency are completed, detect the hardware module state again to smoothly transition to the target state.

[0013] The present invention also provides a power state adaptive switching system for an intelligent terminal device, and the system includes: A hardware characteristic analysis module, which is used to obtain the characteristics of different hardware modules of the intelligent terminal device, determine the multi-stage power state switching process of each hardware module, and obtain the specific control steps and parameter adjustment ranges of the pre-switching, mid-switching, and switching completion stages of each hardware module; An adaptive switching module, which is used to set a local power management state adaptive switching strategy; A switching execution module, when performing power management state adaptive switching, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges of the pre-switching, mid-switching, and switching completion stages of each module, adopts a mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

[0014] The technical solution provided by the present invention has the following beneficial effects: The present invention discloses a method and system for adaptively switching the power state of an intelligent terminal device. The method determines the control steps and parameter adjustment ranges in three stages: pre-switching, in-switching, and post-switching by analyzing the hardware characteristics. In the pre-switching stage, the parameter change amplitude is calculated and an adjustment plan is generated. In the in-switching stage, a progressive strategy is adopted to achieve a smooth transition through multiple small adjustments. At the same time, the power consumption change is monitored in real time, compared with a preset threshold, and the strategy is dynamically optimized to ensure that the power consumption fluctuation is controllable. In the post-switching stage, the hardware module status is detected in multiple rounds to determine whether the target state is reached. Through data accumulation and strategy iteration, the present invention establishes a switching strategy model adapted to different hardware characteristics, realizes the adaptive optimization control of the power state switching process, minimizes the power consumption fluctuation during the switching process, and improves the accuracy and stability of the power management of the hardware module. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 FIG. is a flowchart of a method for adaptively switching the power state of an intelligent terminal device provided by an embodiment of the present invention.

[0016] Figure 2 FIG. is a schematic diagram of another method for adaptively switching the power state of an intelligent terminal device provided by an embodiment of the present invention.

[0017] Figure 3 FIG. is a schematic structural diagram of another system for adaptively switching the power state of an intelligent terminal device provided by an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The technical solutions in the embodiments of the present invention will be clearly and detailedly described below with reference to the accompanying drawings in the embodiments of the present invention. The described embodiments are only a part of the embodiments of the present invention.

[0019] As Figures 1-3 , an embodiment of the present invention provides a method for adaptively switching the power state of an intelligent terminal device, which may specifically include: S100. Obtain the characteristics of different hardware modules of the intelligent terminal device, determine the multi-stage power state switching process of each hardware module, and obtain the specific control steps and parameter adjustment ranges in three stages: pre-switching, in-switching, and post-switching for each hardware module; S110. Set the local power management state adaptive switching strategy; S120. When performing the adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges in three stages: pre-switching, in-switching, and post-switching for each module, adopt a mode switching smooth transition algorithm to achieve a smooth transition of each parameter through progressive adjustment.

[0020] In the above step S100, preferably, obtaining the characteristics of different hardware modules of the intelligent terminal device, determining the multi-stage power state switching process of each hardware module, and obtaining the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching for each hardware module, including: Obtaining characteristic parameters of at least one hardware module, where the characteristic parameters include power consumption and response time; Determining the power state adaptive switching process of the intelligent terminal device according to the characteristic parameters; Determining the duration of each stage; Obtaining power consumption data during the power switching process; Analyzing the power consumption data using a clustering algorithm to obtain the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching for the hardware module. When performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching for each module, using a mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

[0021] In the above step S110, preferably, setting the local power management state adaptive switching strategy includes: Collecting one or more power energy states of the power supply; Comparing with a preset power energy state threshold to determine whether it is lower than the preset power energy state threshold; When the collected power energy state is lower than the preset power energy state threshold, the device enters the low power mode; Processing the historical data sampled in the low power mode, using the processing result as a model input variable to train a neural network model, and feeding back the prediction result output by the obtained neural network model to the energy control system; The energy control system controls the energy switching module to switch the device to the solar power supply mode or the wind power supply mode or the light power supply mode or the electric power supply mode according to the prediction result.

[0022] Optionally, the method further includes: When the collected power energy state is higher than the preset power energy state threshold, entering the high power mode; Processing the historical data sampled in the high power mode, using the processing result as a model input variable to train a neural network model, and feeding back the prediction result output by the obtained neural network model to the energy control system. The energy control system controls the energy switching module to switch the device from the solar power supply mode or the wind power supply mode or the light power supply mode or the electric power supply mode to the intelligent scheduling mode according to the prediction result.

[0023] Optionally, the method further includes: When the collected power state is equal to the preset power state threshold, enter the intelligent scheduling mode; Among them, in the intelligent scheduling mode, the collected solar power generation, wind power generation, light power generation or electric energy is preferentially used for power supply.

[0024] In the above step S120, preferably, when performing the adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching of each module, the mode switching smooth transition algorithm is adopted, and the smooth transition of various parameters is achieved through progressive adjustment, including: When performing the adaptive switching of the power management state, in the pre-switching stage, by analyzing the difference between the hardware module state and the target power state, calculate the change range of the power state parameters that need to be adjusted, and generate a parameter adjustment plan; After entering the in-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, adopt the mode switching smooth transition algorithm, and perform multiple small-amplitude voltage and frequency adjustments through a progressive parameter adjustment strategy to achieve the smooth transition of the hardware module state; Enter the post-switching stage, and judge whether the target power state has been reached by detecting the hardware module state a preset number of times. If the state is stable, officially enter the new power state. If the state is still unstable, return to the in-switching stage to continue fine-tuning the parameters.

[0025] Optionally, the method further includes: When performing parameter adjustment in the in-switching stage, monitor the power consumption change of the hardware module in real time. By comparing with the preset power consumption threshold, if the actual power consumption exceeds the threshold, reduce the adjustment speed or the adjustment amplitude.

[0026] Optionally, when performing the adaptive switching of the power management state, in the pre-switching stage, by analyzing the difference between the hardware module state and the target power state, calculate the change range of the power state parameters that need to be adjusted, and generate a parameter adjustment plan, including: When performing the adaptive switching of the power management state, in the pre-switching stage, obtain the operating state parameters of the current hardware. The operating state parameters include the current voltage and frequency parameters, and form the current hardware module state vector according to the operating state parameters; Obtain the target voltage and target frequency parameters in the target power state, and form the target power state vector according to the target voltage and target frequency parameters; Using the support vector machine algorithm, with the current hardware module state vector as the input and the target power supply state vector as the output, train and form a mapping model from the hardware module state to the power supply state; Input the current hardware module state vector into the mapping model to obtain the adjustment amplitudes of the voltage and frequency parameters required to switch to the target power supply state under the current hardware module state; According to the adjustment amplitudes of the voltage and frequency parameters, formulate a parameter adjustment plan, where the parameter adjustment plan includes the adjustment time points, adjustment step sizes, and adjustment durations of each parameter; Using the decision tree algorithm, with the current values and target values of the voltage and frequency parameters as features and whether the parameters can be adjusted in place as the judgment basis, if the adjustment plan is feasible, initiate pre-switching, and if the adjustment plan is not feasible, adjust the parameter adjustment plan until the adjustment plan is feasible.

[0027] Optionally, after entering the in-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, adopt a mode switching smooth transition algorithm and perform multiple small-amplitude voltage and frequency adjustments through a progressive parameter adjustment strategy to achieve a smooth transition of the hardware module state, including: After entering the in-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, obtain the voltage and frequency parameters that need to be adjusted, as well as the adjustment amplitude and number of times for each adjustment; Judge whether the current hardware module state meets the switching conditions. If it meets, enter the in-switching stage; otherwise, continue to monitor the hardware module state until the switching conditions are met; In the in-switching stage, according to the parameters in the parameter adjustment plan, perform the first small-amplitude adjustment on the voltage, and obtain the current hardware power consumption value after the adjustment; Compare the adjusted power consumption value with the power consumption value before the adjustment, calculate the power consumption fluctuation amplitude, and judge whether it exceeds the preset power consumption fluctuation threshold; If the power consumption fluctuation amplitude does not exceed the threshold, perform the next small-amplitude voltage adjustment according to the parameter adjustment plan until all voltage adjustments are completed; After all voltage adjustments are completed, perform multiple small-amplitude adjustments on the frequency parameters in the same way, and monitor the power consumption fluctuation amplitude after each adjustment; When all progressive adjustments of voltage and frequency are completed, detect the hardware module state again to smoothly transition to the target state.

[0028] Preferably, in the above step S100, obtain the characteristics of different hardware modules, determine the multi-stage power state switching process of each module, and obtain the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching for each module. Specifically, obtain the characteristic parameters of at least one hardware module, where the characteristic parameters include power consumption and response time; determine the multi-stage power state switching process and the duration of each stage according to the characteristic parameters; obtain the power consumption data during the power switching process, and use a clustering algorithm to analyze the power consumption data to obtain the optimal parameter configuration range of the hardware module in each switching stage; apply the optimized multi-stage power switching process and the parameter configuration of each stage to the power management of the hardware module to achieve efficient and low-power operation of the hardware module.

[0029] Exemplarily, in order to obtain characteristic parameters such as power consumption and response time of a hardware module, a measurement module can be used to measure the current and voltage waveforms in different working states, and the power consumption parameters can be calculated by analyzing indicators such as the peak value and steady-state value of the waveforms. For example, if the measured peak current of a module in the working state is 500 mA and the voltage is 3 V, then its peak power consumption is 65 W. By comparing and analyzing the parameters of multiple modules, a reasonable power switching process is determined, such as switching the module with smaller power consumption first, and then gradually switching the module with larger power consumption. The duration of each stage is determined according to the response time of the module, such as setting the switching time to more than twice the response time. In the pre-switching stage, obtain the control steps and parameter ranges according to the process, such as gradually reducing the voltage and clock frequency, with the voltage reduction range being 1 V to 3 V and the frequency reduction range being 10% to 30%. During the in-switching stage, continuously adjust the voltage and frequency to control the module to perform state switching until the target state is reached. After the switching is completed, determine whether the target state has been reached by reading the value of the power status register. During the switching process of multiple modules, analyze the switching time and power consumption of each module through a decision tree algorithm, and dynamically adjust the timing of the process. For example, if the switching time of a certain module exceeds 20% of the expected value, then extend the time of the entire process. Analyze the power consumption data during the switching process through a clustering algorithm to obtain the optimal parameter range of each stage, such as the optimal voltage reduction range in the pre-switching stage being 2 V to 25 V. Finally, apply the optimized switching process and parameter configuration to the hardware module, and achieve automated power management through technologies such as state machines to achieve the purpose of efficient and low-power operation.

[0030] According to the characteristic parameters of the hardware module, use a clustering algorithm to analyze the power consumption data during the power switching process to obtain the optimal parameter configuration range of the hardware module in each switching stage. By applying the optimized multi-stage power switching process and the parameter configuration of each stage to the power management of the hardware module, efficient and low-power operation of the hardware module is achieved.

[0031] Obtain the power consumption data of the hardware module at different power switching stages, and select a suitable clustering algorithm for analysis and processing according to the characteristics of the data. Divide the power consumption data into different clusters through the clustering algorithm, with each cluster representing a power switching stage, and determine the optimal parameter configuration range for this stage according to the distribution of the data within the cluster. Optimize the original power switching process, divide the process into multiple stages, with each stage corresponding to an optimal parameter configuration range, to form an optimized multi-stage power switching process. Apply the optimized multi-stage power switching process to the power management of the hardware module, and dynamically adjust the parameter configuration according to the current switching stage to keep it always within the optimal range. Continuously collect power consumption data during the operation of the hardware module, use the incremental learning algorithm to update and optimize the clustering results, and timely adjust the parameter configuration ranges of each stage. If it is found that the current parameter configuration deviates from the optimal range, trigger the adjustment of the power management strategy, re-search for the optimal parameters, and switch the operating state of the hardware module to the new parameter configuration. Through continuous data collection, analysis, and parameter optimization, the hardware module can maintain efficient and low-power operation at each power switching stage, improving the overall energy efficiency performance.

[0032] Exemplarily, first, power consumption data of the hardware module at different power supply switching stages is collected by a sensor, once every 10 milliseconds for 1 minute continuously, obtaining a total of 6000 data points. According to the characteristics of the data, the K-means clustering algorithm is selected for analysis and processing. The optimal number of clusters is determined to be 4 through the Elbow Method, and the power consumption data is divided into 4 clusters, each cluster corresponding to 4 power supply switching stages of low-power standby, normal operation, high-performance operation, and overload protection respectively. The Elbow Method is a technique used to determine the best number of clusters in cluster analysis. According to the distribution of the data within the clusters, the Gaussian mixture model is used to model the data of each cluster, obtaining the mean and variance of the power consumption data at each stage, and determining the optimal parameter configuration range for this stage with μ±96σ. For example, the voltage range in the low-power standby stage is 3V±1V, and the clock frequency range is 12MHz±5MHz. The original power supply switching process is optimized by introducing the fuzzy control theory. According to the matching degree between the current power consumption level and the optimal range at each stage, the voltage and clock frequency parameters are smoothly adjusted to form an optimized multi-stage power supply switching process. The optimized multi-stage power supply switching process is applied to the power management of the hardware module. Using the predictive control algorithm, according to the historical power consumption data, the power consumption change trend in the next period of time is predicted, and the parameter configuration is adjusted in advance to keep it always within the optimal range. During the operation of the hardware module, power consumption data is continuously collected at intervals of 10 milliseconds, and the incremental learning algorithm is used to update and optimize the clustering results in real time. When the distance between the newly collected data point and the original clustering center exceeds the threshold, reclustering is triggered, and the parameter configuration range at each stage is adjusted accordingly. If it is found that the current parameter configuration deviates from the optimal range by more than 5%, the adjustment of the power management strategy is triggered, and the particle swarm optimization algorithm is used to search for the optimal parameter combination, and the operating state of the hardware module is smoothly switched to the new parameter configuration within 10 milliseconds. Through continuous data collection, analysis, and parameter optimization, the hardware module can maintain efficient and low-power operation at each power supply switching stage, with a 15% increase in power efficiency and a 12% reduction in power consumption.

[0033] Preferably, in the above step S110, one or more power states of the electrical equipment are collected; the intelligent terminal device first collects the power state information of the environment in real time, and this information may include but is not limited to battery level, light intensity, wind speed or solar radiation level. According to the one or more power state information collected, the device intelligently determines the power management mode that matches the current environmental conditions. For example, when the battery level is lower than the preset threshold, the device will automatically enter the low battery mode and use historical data to train a neural network model to predict the future energy situation. The prediction results guide the energy control system to select different power supply modes such as solar energy, wind energy, light energy or electrical energy. On the contrary, when the power state is higher than the preset threshold, the device enters the high battery mode and also uses historical data to optimize power management. This method enables the device to adaptively adjust the power management mode, effectively improving the energy utilization efficiency and the working reliability of the device. In the low battery mode and the high battery mode, the historical data is preprocessed, including data cleaning, missing value filling, outlier handling, etc. Key features are extracted, such as the battery level change rate, device power consumption pattern, environmental conditions, etc. These features will be used as input variables for the neural network model. The processing results are used as model input variables to train the neural network model. Neural network models such as LSTM (Long Short-Term Memory) or GRU (Gated Recurrent Unit) are used to capture the dynamic power state sequence. The model should include an input layer, a hidden layer and an output layer, where the hidden layer can capture the patterns of time series data. LSTM and GRU are two commonly used time series prediction models, and they both have good long-term and short-term memory capabilities and are suitable for processing long sequence data with strong time correlations. By constructing an LSTM or GRU model, the energy change law over time can be learned. In the low battery mode, the prediction results output by the obtained neural network model are fed back to the energy control system, and the energy control system controls the energy switching module to switch the device to the solar power supply mode or the wind power supply mode or the light power supply mode or the electrical power supply mode according to the prediction results; when the collected energy state is higher than the preset threshold, it enters the high battery mode. At this time, the historical data sampled in the high battery mode is processed. In the low battery mode and the high battery mode, the historical data is preprocessed, including data cleaning, missing value filling, outlier handling, etc.Extract key features such as the rate of power change, device power consumption mode, environmental conditions, etc. These features will serve as input variables for the neural network model. The processing results will be used as input variables to train the neural network model. The predicted results output by the obtained neural network model will be fed back to the energy control system. The energy control system will control the energy switching module to switch the device from the solar power supply mode, wind power supply mode, light power supply mode, or electric power supply mode to the intelligent scheduling mode according to the predicted results; if it equals the preset threshold, it will enter the intelligent scheduling mode; among them, in the intelligent scheduling mode, the collected solar power generation, wind power generation, light power generation, and electric power are preferentially used for power supply, and the best choice is made according to the advantages and disadvantages of each item.

[0034] Taking a mobile phone as an example, the acquisition of the electrical energy state can be the acquisition of the instantaneous voltage and instantaneous current of the power supply. The acquisition of the electrical energy state can obtain the instantaneous voltage and instantaneous current of the power supply in real time, so as to accurately reflect the current working condition of the power supply, provide key data support for subsequent power management and energy efficiency optimization, help users discover and solve abnormal power consumption in time, reduce energy waste, and improve the operation efficiency and safety of the device. Connect the device to the power line, obtain the voltage and current data of the power supply through a specific interface, and the data is processed by the system built in the device and the results are output. The user can view the power supply status through the display screen or remote communication device. The sensors installed in the device are responsible for detecting the voltage and current of the power supply in real time and transmitting the data to the core processor of the device. The processor analyzes the data through algorithms, judges the current power supply status, and adjusts the power supply output or issues an alarm according to the preset program. The user can set the alarm threshold independently to respond to different types of power supply status changes.

[0035] Furthermore, a power state detection module, an ADC module and a CPU module connected thereto are provided on the main control microprocessor of the mobile phone. The power state detection module includes a battery power detection module, a WiFi communication module and / or a Bluetooth communication module, which can realize the comprehensive monitoring of the power state of the mobile phone. Specifically, this design can detect the remaining battery power of the mobile phone in real time and transmit the battery power information to the CPU module through the WiFi communication module or the Bluetooth communication module, so that the CPU module can effectively adjust the working mode of the mobile phone according to the real-time power state, such as reducing unnecessary data transmission, lowering the screen brightness, etc., to achieve the effects of energy saving and extending the battery life. At the same time, the key data such as the battery voltage is accurately collected through the ADC module, further enhancing the accuracy of the power state detection and ensuring the stability and reliability of the mobile phone in various usage scenarios.

[0036] In the low - power mode or high - power mode, perform smoothing filtering processing and historical data storage processing on the collected voltage, current, and power data; based on the obtained stable historical data, simulate and analyze the future power state, calculate different switching time points and corresponding possible performance values, thereby helping users better manage power, improve power utilization efficiency, and ensure the stability and reliability of the power supply system.

[0037] For example, on the main control micro - processor of a mobile phone, there are also a GPS positioning module, a temperature detection module, an audio sensor, a motion sensor, touch - screen brightness adjustment, and a camera. When the power state of the system is equal to the preset power state threshold, it will automatically enter the intelligent scheduling mode. In this mode, it will preferentially use the collected solar power generation, wind power generation, light power generation, and electric power for power supply, and make a choice according to the advantages and disadvantages of different energy sources. At the same time, components such as the GPS positioning module, temperature detection module, audio sensor, motion sensor, touch - screen brightness adjustment function, and camera integrated in the main control micro - processor of the mobile phone can, based on the results of the above - mentioned intelligent scheduling selection, achieve more accurate and efficient energy management and device control, thereby optimizing the energy utilization efficiency of the mobile phone, enhancing the user experience, and ensuring stable and reliable power supply in different environments. Among them, the GPS positioning module includes a three - in - one module of a GPS satellite navigation chip, an inertial navigation chip, and an electronic compass chip.

[0038] Optionally, in the pre - switching stage of step S120, by analyzing the differences between the current hardware module state and the target power state, calculate the change amplitudes of parameters such as voltage and frequency that need to be adjusted, and generate a detailed parameter adjustment plan to prepare for the subsequent switching.

[0039] S121. Obtain the operating state parameters of the current hardware; Among them, the operating state parameters include the current voltage and frequency parameters; S122. Form a current hardware module state vector according to the operating state parameters; S123. Obtain the target voltage and target frequency parameters in the target power state; S124. Form a target power state vector according to the target voltage and target frequency parameters; S125. Adopt the support vector machine algorithm, use the current hardware module state vector as the input and the target power state vector as the output, and train to form a mapping model from the hardware module state to the power state; S126. Input the current hardware module state vector into the mapping model to obtain the adjustment amplitudes of the voltage and frequency parameters required to switch to the target power state under the current hardware module state; S127. Develop a parameter adjustment plan according to the adjustment ranges of the voltage and frequency parameters; Among them, the parameter adjustment plan includes the adjustment time points, adjustment step sizes, and adjustment durations of each parameter; S128. Use a decision tree algorithm to judge and adjust the plan; Among them, use a decision tree algorithm, with the current values and target values of the voltage and frequency parameters as features, and whether the parameters can be adjusted in place as the judgment basis. If the adjustment plan is feasible, initiate pre-switching. If the adjustment plan is not feasible, adjust the parameter adjustment plan until the adjustment plan is feasible.

[0040] Exemplarily, first collect the operating state parameters of the hardware through sensors, such as the current voltage is 2V, the frequency is 4GHz, etc., to form a 10-dimensional hardware module state vector. Then obtain the target power supply state, such as the target voltage 8V, frequency 2GHz, etc. parameters in the standby state, to form an 8-dimensional target state vector. Next, use the SVM support vector machine algorithm, with the hardware module state vector as the input and the target state vector as the output, and train a mapping model through the Gaussian kernel function. Input the current hardware module state vector into this model, and a series of adjustment ranges such as the voltage needs to be reduced from 2V to 9V, the frequency needs to be reduced from 4GHz to 8GHz, etc. can be obtained. Develop a detailed parameter adjustment plan according to the adjustment ranges, such as the voltage drops 0.5V every 5ms and lasts for 60ms; the frequency drops 1GHz every 10ms and lasts for 60ms, etc. Then use the CART decision tree algorithm, with the current voltage 2V, target voltage 9V, etc. as features, and whether the voltage adjustment is in place as the judgment basis, to generate a feasibility prediction model. Input the adjustment plan into this model for prediction. If it is feasible, initiate pre-switching. If it is not feasible, such as the voltage cannot drop to 9V within 60ms, the parameter plan needs to be adjusted, such as changing the drop step size to 0.3V every 5ms and increasing the duration to 100ms until the prediction is feasible. Through the above process, a smooth transition from the hardware module state to the power supply state can be achieved.

[0041] According to the current hardware operating state parameters and target power supply state parameters, use the support vector machine algorithm to train and form a mapping model from the hardware module state to the power supply state. Through this model, obtain the adjustment ranges of the voltage and frequency parameters, and accordingly develop a parameter adjustment plan to determine the adjustment time points, adjustment step sizes, and durations of each parameter.

[0042] Optionally in the above step S120, after entering the switching-in stage, according to the adjustment plan generated in the pre-switching stage, adopt a progressive parameter adjustment strategy to achieve a smooth transition of the hardware module state through multiple small-amplitude voltage and frequency adjustments, and avoid power consumption fluctuations caused by large-amplitude parameter changes.

[0043] According to the adjustment plan generated in the pre-switching stage, obtain the voltage and frequency parameters to be adjusted, as well as the amplitude and number of times of each adjustment. Determine whether the current hardware module status meets the switching conditions. If it meets, enter the in-switching stage; otherwise, continue to monitor the hardware module status until the switching conditions are met. In the in-switching stage, according to the parameters in the adjustment plan, make a first small adjustment to the voltage. After the adjustment, obtain the current hardware power consumption value. Compare the adjusted power consumption value with the power consumption value before the adjustment, calculate the power consumption fluctuation amplitude, and determine whether it exceeds the preset power consumption fluctuation threshold. If the power consumption fluctuation amplitude does not exceed the threshold, make the next small voltage adjustment according to the adjustment plan, and repeat the steps until all voltage adjustments are completed. After all voltage adjustments are completed, make multiple small adjustments to the frequency parameters in the same way, and monitor the power consumption fluctuation amplitude after each adjustment. When all progressive adjustments of voltage and frequency are completed, detect the hardware module status again to ensure a smooth transition to the target state and complete the in-switching stage.

[0044] Exemplarily, according to the adjustment plan generated in the pre-switching stage, the voltage needs to be adjusted from 2V to 0V, with an adjustment of 0.5V each time for a total of 4 adjustments; the frequency needs to be adjusted from 4GHz to 8GHz, with an adjustment of 1GHz each time for a total of 6 adjustments. By detecting key indicators such as CPU occupancy rate and memory usage rate, it is determined that the current hardware module status meets the switching conditions and enters the in-switching stage. First, adjust the voltage to 1.5V, and use a power meter to obtain a power consumption of 10W. Compared with the power consumption of 12W before the adjustment, the fluctuation amplitude is 17%, which does not exceed the preset threshold of 20%. Continue the next adjustment to 1.0V, and the power consumption becomes 9W, with a fluctuation amplitude of 10%. And so on, until the voltage is adjusted to 0V and the power consumption stabilizes at about 8W. Then, adjust the frequency to 8GHz step by step in the same way, and monitor the power consumption fluctuation amplitude after each adjustment to ensure that it is within the threshold. Finally, detect again that the CPU occupancy rate drops below 60% and the memory usage rate drops below 70%, confirming that the target state has been reached and the switching is completed smoothly. The whole process is realized through an automated script, which dynamically adjusts the voltage and frequency according to the hardware module status and power consumption feedback, without manual intervention, ensuring the stability and efficiency of the switching.

[0045] By obtaining the hardware module status parameters, determining whether the preset switching condition threshold is met, making progressive small adjustments to the voltage and frequency according to the adjustment plan, obtaining the power consumption fluctuation amplitude after each adjustment, and determining that a smooth transition to the target state has been achieved after the adjustment.

[0046] Optionally, in the above step S120, during the parameter adjustment process, real-time monitor the power consumption change of the hardware. By comparing with the preset power consumption threshold, dynamically optimize the adjustment strategy. If the actual power consumption exceeds the threshold, slow down the adjustment speed or reduce the adjustment amplitude to ensure that the power consumption fluctuation is within a controllable range.

[0047] Obtain the real-time power consumption data of the hardware, compare it with the preset power consumption threshold, and get the difference between the actual power consumption and the threshold. According to the difference between the actual power consumption and the threshold, adopt a fuzzy control algorithm to dynamically adjust the adjustment strategy of the hardware parameters, and determine the appropriate adjustment speed and adjustment amplitude. If the actual power consumption exceeds the threshold, judge whether it is necessary to reduce the adjustment speed or decrease the adjustment amplitude through fuzzy inference rules to avoid excessive power consumption fluctuations. During the process of adjusting the hardware parameters, continuously monitor the change of the hardware power consumption, obtain the real-time power consumption data, and use it as the input for the optimization of the next adjustment strategy. Through the support vector machine algorithm, train the historical adjustment strategy and the corresponding power consumption fluctuation situation to obtain the optimal adjustment strategy model, which is used to guide the subsequent parameter adjustment. Apply the optimized adjustment strategy to the process of adjusting the hardware parameters, and monitor its impact on the power consumption fluctuation in real time to ensure that the power consumption is always within the controllable range. According to the long-term change trend of the hardware power consumption, adopt an adaptive learning algorithm to dynamically update the preset power consumption threshold to adapt to the change of the hardware performance and ensure the effectiveness of the adjustment strategy.

[0048] Exemplarily, the hardware power consumption data is collected in real time through sensors, such as CPU utilization rate, memory occupancy, etc., and sampled every 5 seconds. Calculate the difference between the collected data and the preset threshold (such as CPU utilization rate of 80% and memory occupancy of 70%) to obtain the deviation degree, which is used as the input of the fuzzy control algorithm. The fuzzy rule base sets 9 rules, and outputs the adjustment speed (-5% to +5%) and amplitude (-10% to +10%) according to the deviation degree. If the CPU utilization rate exceeds the threshold by 10%, reduce the adjustment speed by 20% and decrease the adjustment amplitude by 30%. The SVM algorithm trains the optimal model based on the historical 30 groups of adjustment strategies and power consumption data, and the accuracy rate reaches 90%. The adaptive learning algorithm dynamically adjusts the CPU utilization rate threshold from 80% to 85% and the memory occupancy threshold from 70% to 75% based on the power consumption change trend in the recent 24 hours to adapt to the business peak period.

[0049] According to the comparison of the difference between the real-time power consumption data and the preset threshold, adopt a fuzzy control algorithm to dynamically adjust the hardware parameters, judge whether it is necessary to reduce the adjustment speed or decrease the adjustment amplitude through fuzzy inference rules to avoid excessive power consumption fluctuations, and determine the appropriate adjustment strategy.

[0050] Optionally, in the above step S120, after all parameter adjustments are completed, enter the switching completion stage. Through multiple rounds of detection of the state of the hardware module, judge whether the target power state has been reached. If the state is stable, officially enter the new power state. If the state is still unstable, return to the switching stage to continue with fine-tuning of the parameters.

[0051] According to the preset target power supply state, obtain the operating parameters of the current hardware, compare them with the target parameters, and determine the parameter items that need to be adjusted. For the parameter items that need to be adjusted, use the gradient descent algorithm, and through multiple iterative calculations, obtain the optimal parameter adjustment value. Apply the calculated parameter adjustment value to the hardware system, send instructions through the controller, and finely adjust the parameters of the hardware. After the parameter adjustment is completed, enter the switching completion stage, obtain the real-time operating state data of the hardware, and use the Kalman filter algorithm to filter the data to obtain a stable state evaluation value. Compare the state evaluation value with the preset stable state threshold. If the evaluation value is within the threshold range, it is determined that the state of the hardware module is stable, and it officially enters the new power supply state. If the evaluation value exceeds the stable threshold range, it is considered that the state of the hardware module is still unstable, return to the switching stage, and use the evolutionary algorithm to finely adjust and optimize the parameters according to the state deviation value. Repeat the steps until the state of the hardware module is stable in the target power supply state, complete the power supply switching process, and write the switching result data into the system log for subsequent analysis and optimization.

[0052] Exemplarily, according to a preset target power supply state, the system obtains the operating parameters of the current hardware through sensors, such as a voltage of 220V, a current of 5A, a power of 330W, etc., and compares them with the target parameters such as a voltage of 220V, a current of 2A, and a power of 264W to determine that the parameter items to be adjusted are the current and power. For the current and power parameters to be adjusted, the system uses the gradient descent algorithm, sets the learning rate to 01, and the number of iterations to 100 times. Through multiple iterative calculations, the optimal parameter adjustment values are obtained, that is, the current is adjusted to 2A and the power is adjusted to 264W. Apply the calculated parameter adjustment values to the hardware system, and send a standard current command of 4-20mA through the PLC controller to finely adjust the current and power parameters of the hardware. After the parameter adjustment is completed, enter the switching completion stage. The system obtains the real-time operating status data of the hardware every 1s, uses the Kalman filter algorithm to filter the data, sets the process noise covariance to 01, the measurement noise covariance to 1, and obtains a stable state evaluation value through a first-order linear recurrence equation. Compare the state evaluation value with the preset stable state threshold. If the evaluation value is within the threshold range (such as the voltage error is less than 1V, the current error is less than 1A, and the power error is less than 5W), it is determined that the hardware module state is stable and officially enters the new power supply state. If the evaluation value exceeds the stable threshold range, it is considered that the hardware module state is still unstable. The system automatically returns to the switching stage, and according to the state deviation value, uses the particle swarm optimization algorithm to finely adjust and optimize the parameters again. Set the number of particles to 50, the inertia weight to 8, the acceleration constant to 5, and the number of iterations to 80 times. Continuously update the position and velocity of the particles to search for the optimal parameter adjustment value. Repeat the above steps until the hardware module state is stable in the target power supply state, complete the power supply switching process, and write the switching result data into the system log database. Use the decision tree algorithm to classify and analyze the log data, continuously optimize the control strategy and switching process, and improve the reliability and efficiency of the system.

[0053] Send instructions through the controller to finely adjust the parameters of the hardware, use the Kalman filter algorithm to filter the data to obtain a stable state evaluation value, and use the evolutionary algorithm to finely adjust and optimize the parameters again according to the state deviation value.

[0054] The controller obtains the current operating parameters of the hardware and transmits the parameter data to the algorithm module. The algorithm module uses the Kalman filter algorithm to filter the received parameter data, removing the noise interference in the data and obtaining a stable state evaluation value. According to a preset threshold, it judges whether the state evaluation value exceeds the normal range. If it exceeds, it determines that there is a state deviation and triggers the evolutionary algorithm. The evolutionary algorithm takes the state deviation value as the optimization goal and searches for the optimal fine-tuning parameter combination through iterative evolution. The optimal fine-tuning parameter combination obtained by the evolutionary algorithm is sent to the controller, and the controller performs fine-tuning control on the hardware according to the received parameters. After the hardware performs fine-tuning, the controller obtains the hardware operating parameters again and transmits them to the algorithm module for a new round of state evaluation. Repeat the steps, and continuously optimize the hardware operating state through closed-loop feedback control until the state deviation value converges within the preset threshold, achieving precise control of the hardware.

[0055] Exemplarily, the controller obtains the current operating parameters of the hardware, including temperature, pressure, flow rate, etc., through the RS485 bus, and transmits the parameter data to the algorithm module at a frequency of 10 times per second. The algorithm module uses the Kalman filter algorithm to filter the received parameter data, sets the measurement noise variance to 1 and the process noise variance to 0.1, and removes the Gaussian white noise interference in the data through 5 iterations to obtain a stable state evaluation value. According to the preset normal range threshold, it judges whether the state evaluation value exceeds the normal range. For example, the normal range of temperature is 20°C - 80°C. If it exceeds, it determines that there is a state deviation and triggers the differential evolution algorithm. The evolutionary algorithm takes the state deviation value as the optimization goal, sets the population size to 50, the crossover probability to 8, and the mutation probability to 1, and searches for the optimal control parameter combination through 100 iterations of evolution. The optimal parameter combination obtained by the evolutionary algorithm is sent to the controller, and the controller performs fine-tuning control on the actuators of the hardware, such as motors, valves, etc., according to the received parameters. After the hardware performs fine-tuning, the controller obtains the hardware operating parameters again and transmits them to the algorithm module for a new round of state evaluation. Continuously optimize the hardware operating state through closed-loop feedback control, set the deviation convergence threshold to 1%, and repeat the iteration until the state deviation value converges within the threshold, achieving precise control of the hardware and ensuring the long-term stable operation of the system.

[0056] Optionally, in the above step S120, the optimized switching strategy is applied to subsequent power state switching. Through continuous data accumulation and strategy iteration, a switching strategy model adapted to different hardware characteristics is established to achieve adaptive optimization control of the power state switching process and minimize the power consumption fluctuation during the switching process.

[0057] Obtain real-time power consumption data during the power state switching process, and judge the pros and cons of the current switching strategy according to the change trend of the power consumption data. If the current switching strategy causes large power consumption fluctuations, trigger the update and optimization process of the strategy model. Use machine learning algorithms to obtain better switching strategy parameters through learning and training of historical power consumption data. Write the optimized switching strategy parameters into the strategy model to form a new switching strategy. During subsequent power state switching processes, use the updated switching strategy for control. During the power state switching process, continuously monitor the change of power consumption data. If it is detected that the power consumption fluctuation amplitude exceeds the preset threshold, trigger the alarm mechanism to prompt that the switching strategy needs to be further optimized. According to the characteristic parameters of different hardware, select the most matching switching strategy model from the strategy model library to guide the power state switching of this hardware. Regularly collect power state switching data of different hardware through the cloud server, and use big data analysis technology to mine the power consumption characteristics of various hardware, continuously enriching and improving the strategy model library. The strategy model library stores multiple switching strategy models, and each model corresponds to a specific hardware type. The system adaptively selects the corresponding switching strategy model according to the recognition result of the hardware type to achieve Exemplarily, during the power state switching process, the system collects power consumption data every 10 milliseconds and transmits the data to the edge computing node. The node uses the Kalman filter algorithm to filter the power consumption data, filter out high-frequency noise interference, and extract the power consumption change trend. If within 100 milliseconds, the power consumption fluctuation amplitude exceeds 5% of the rated power consumption, it is determined that the current switching strategy is not optimal, and the strategy model update process is triggered. The system randomly extracts 1000 groups of data from the historical power consumption database, and each group of data contains the power consumption sampling values within 1 second before and after the switch. Using the long short-term memory neural network (LSTM) algorithm, the data is trained and learned for 500 iterations to obtain the optimized switching strategy parameters. The updated strategy parameters are written into the strategy model and deployed to the embedded controller. The controller dynamically adjusts the power state switching timing according to the new strategy model to smooth the power consumption fluctuation. If during the actual operation of the optimized strategy, the power consumption fluctuation exceeds 8% of the rated power consumption, the system automatically sends an alarm message to the cloud management platform to notify the administrator to optimize the strategy. The cloud management platform semantically models the characteristic parameters of different hardware based on the hardware resource description framework to form a hardware knowledge graph. Through the graph inference engine, the hardware type most similar to the target hardware is identified, and the corresponding switching strategy model is selected from the strategy model library. The strategy model library is stored in a database, supporting multi-dimensional strategy model retrieval and association analysis. The system regularly collects the power state switching logs from each hardware node every day and uploads them to the cloud. The cloud uses a big data processing framework to clean and normalize the log data. Then, the association rule mining algorithm is used to discover the power consumption characteristics and laws of different hardware types, and the mining results are fed back to the strategy model library to continuously optimize and enrich the model.

[0058] By obtaining the real-time power consumption data during the power state switching process, judging the pros and cons of the current switching strategy according to the power consumption change trend, using machine learning algorithms to obtain more optimal switching strategy parameters, forming a new switching strategy model, storing it in the strategy model library, and adaptively selecting the most matching strategy model according to the hardware characteristic parameters to guide the power state switching of the hardware and achieve the effect of power consumption optimization.

[0059] An embodiment of the present invention further provides a power state adaptive switching system for an intelligent terminal device, and the system includes: A hardware characteristic analysis module, configured to obtain the characteristics of different hardware modules of the intelligent terminal device, determine the multi-stage power state switching process of each hardware module, and obtain the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching of each hardware module; An adaptive switching module, configured to set a local power management state adaptive switching strategy; The switching execution module, when performing adaptive switching of the power management state, adopts a mode switching smooth transition algorithm according to the multi-stage power state switching processes of each hardware module and the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching, and post-switching for each module, and achieves a smooth transition of various parameters through progressive adjustment.

[0060] Obviously, those skilled in the art can make various changes and modifications to the embodiments of the present application without departing from the spirit and scope of the embodiments of the present application. Thus, if these modifications and variations of the embodiments of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these changes and modifications.

Claims

1. A method for adaptively switching the power state of an intelligent terminal device, characterized in that, Including: Obtain the characteristics of different hardware modules of the intelligent terminal device, determine the multi-stage power state switching process of each hardware module, and obtain the specific control steps and parameter adjustment ranges for the three stages of pre-switching, in-switching, and post-switching of each hardware module; Set the local power management state adaptive switching strategy; When performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges for the three stages of pre-switching, in-switching, and post-switching of each module, adopt the mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

2. The power state adaptive switching method of the intelligent terminal device according to claim 1, wherein The obtaining the characteristics of different hardware modules of the intelligent terminal device, determining the multi-stage power state switching process of each hardware module, and obtaining the specific control steps and parameter adjustment ranges for the three stages of pre-switching, in-switching, and post-switching of each hardware module includes: Obtain the characteristic parameters of at least one hardware module, where the characteristic parameters include power consumption and response time; According to the characteristic parameters, determine the power state adaptive switching process of the intelligent terminal device and the duration of each stage; Obtain the power consumption data during the power switching process, analyze the power consumption data using a clustering algorithm, and obtain the specific control steps and parameter adjustment ranges for the three stages of pre-switching, in-switching, and post-switching of the hardware module, so that when performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges for the three stages of pre-switching, in-switching, and post-switching of each module, adopt the mode switching smooth transition algorithm to achieve a smooth transition of various parameters through progressive adjustment.

3. The power state adaptive switching method for the intelligent terminal device according to claim 1, wherein, The setting the local power management state adaptive switching strategy includes: Collect one or more power energy states of the power supply; Compare with a preset power energy state threshold to determine whether it is lower than the preset power energy state threshold; When the collected power energy state is lower than the preset power energy state threshold, the device enters the low power mode; Process the historical data sampled in the low power mode, use the processing result as the model input variable to train the neural network model, and feed back the prediction result output by the obtained neural network model to the energy control system; The energy control system controls the energy switching module to switch the device to the solar power supply mode, or the wind power supply mode, or the light power supply mode, or the electric power supply mode according to the prediction result.

4. The power state adaptive switching method for the intelligent terminal device according to claim 3, wherein Also included: When the collected power energy state is higher than the preset power energy state threshold, enter the high power mode; Process the historical data sampled in the high power mode, use the processing result as the model input variable to train the neural network model, and feed back the prediction result output by the obtained neural network model to the energy control system, and the energy control system controls the energy switching module to switch the device from the solar power supply mode, or the wind power supply mode, or the light power supply mode, or the electric power supply mode to the intelligent scheduling mode according to the prediction result.

5. The power state adaptive switching method for the intelligent terminal device according to claim 4, characterized in that Also included: When the collected power energy state is equal to the preset power energy state threshold, enter the intelligent scheduling mode; Among them, in the intelligent scheduling mode, the collected solar power generation, wind power generation, light power generation or electric energy is preferentially used for power supply.

6. The power state adaptive switching method of the intelligent terminal device according to claim 1, characterized in that, When performing adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges in the three stages of pre-switching, in-switching and post-switching of each module, a mode switching smooth transition algorithm is adopted, and through progressive adjustment, a smooth transition of various parameters is achieved, including: When performing adaptive switching of the power management state, in the pre-switching stage, by analyzing the difference between the hardware module state and the target power state, the change amplitude of the power state parameters that need to be adjusted is calculated, and a parameter adjustment plan is generated; After entering the in-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, a mode switching smooth transition algorithm is adopted, and through a progressive parameter adjustment strategy, multiple small-amplitude voltage and frequency adjustments are performed to achieve a smooth transition of the hardware module state; After entering the post-switching stage, through a preset number of detections of the hardware module state, it is judged whether the target power state has been reached. If the state is stable, it officially enters the new power state. If the state is still unstable, it returns to the in-switching stage to continue fine-tuning the parameters.

7. The power state adaptive switching method for the intelligent terminal device according to claim 6, characterized in that It also includes: When performing parameter adjustment in the in-switching stage, the power consumption change of the hardware module is monitored in real time. By comparing with a preset power consumption threshold, if the actual power consumption exceeds the threshold, the adjustment speed or the adjustment amplitude is reduced.

8. The power state adaptive switching method for the intelligent terminal device according to claim 7, wherein When performing adaptive switching of the power management state, in the pre-switching stage, by analyzing the difference between the hardware module state and the target power state, the change amplitude of the power state parameters that need to be adjusted is calculated, and a parameter adjustment plan is generated, including: When performing adaptive switching of the power management state, in the pre-switching stage, the operating state parameters of the current hardware are obtained. The operating state parameters include the current voltage and frequency parameters, and a current hardware module state vector is formed according to the operating state parameters; The target voltage and target frequency parameters in the target power state are obtained, and a target power state vector is formed according to the target voltage and target frequency parameters; Using the support vector machine algorithm, with the current hardware module state vector as the input and the target power state vector as the output, a mapping model from the hardware module state to the power state is trained; The current hardware module state vector is input into the mapping model to obtain the adjustment amplitude of the voltage and frequency parameters required to switch to the target power state in the current hardware module state; According to the adjustment amplitude of the voltage and frequency parameters, a parameter adjustment plan is formulated. The parameter adjustment plan includes the adjustment time point, adjustment step size and adjustment duration of each parameter; Using the decision tree algorithm, with the current values and target values of the voltage and frequency parameters as features and whether the parameters can be adjusted in place as the judgment basis, if the adjustment plan is feasible, the pre-switching is started. If the adjustment plan is not feasible, the parameter adjustment plan is adjusted until the adjustment plan is feasible.

9. The power state adaptive switching method for the intelligent terminal device according to claim 8, wherein After entering the mid-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, the mode switching smooth transition algorithm is adopted, and multiple small-amplitude voltage and frequency adjustments are carried out through a progressive parameter adjustment strategy to achieve the smooth transition of the hardware module state, including: After entering the mid-switching stage, according to the parameter adjustment plan generated in the pre-switching stage, obtain the voltage and frequency parameters to be adjusted, as well as the amplitude and number of each adjustment; Judge whether the current hardware module state meets the switching conditions. If it meets, enter the mid-switching stage; otherwise, continue to monitor the hardware module state until the switching conditions are met; In the mid-switching stage, according to the parameters in the parameter adjustment plan, make the first small-amplitude voltage adjustment, and obtain the current hardware power consumption value after the adjustment; Compare the adjusted power consumption value with the power consumption value before the adjustment, calculate the power consumption fluctuation amplitude, and judge whether it exceeds the preset power consumption fluctuation threshold; If the power consumption fluctuation amplitude does not exceed the threshold, make the next small-amplitude voltage adjustment according to the parameter adjustment plan until all voltage adjustments are completed; After all voltage adjustments are completed, make multiple small-amplitude adjustments to the frequency parameters in the same way, and monitor the power consumption fluctuation amplitude after each adjustment; When all progressive adjustments of voltage and frequency are completed, detect the hardware module state again to smoothly transition to the target state.

10. A power state adaptive switching system for an intelligent terminal device, characterized in that, The system includes: A hardware characteristic analysis module, which is used to obtain the characteristics of different hardware modules of the intelligent terminal device, determine the multi-stage power state switching process of each hardware module, and obtain the specific control steps and parameter adjustment ranges of the pre-switching, mid-switching, and post-switching stages of each hardware module; An adaptive switching module, which is used to set the adaptive switching strategy of the local power management state; A switching execution module, when performing the adaptive switching of the power management state, according to the multi-stage power state switching process of each hardware module and the specific control steps and parameter adjustment ranges of the pre-switching, mid-switching, and post-switching stages of each module, adopts the mode switching smooth transition algorithm to achieve the smooth transition of various parameters through progressive adjustment.

Images