Self-adaptive power supply management and control method and device for self-powered intelligent door lock

Abstract

The invention provides a self-adaptive power supply management and control method and device for a self-powered intelligent door lock, and relates to the technical field of door lock power supply management and control. Performing classification identification and training fusion on the data set by using a recurrent neural network, and constructing an electric quantity consumption adaptive prediction network; acquiring the door lock electric quantity and the working state to perform electric quantity consumption prediction in a preset time window; constructing a power supply control strategy library, matching with the strategy library based on the electric quantity prediction consumption parameter and the current electric quantity state, and determining a power supply control strategy; the door lock electric quantity prediction consumption parameters are optimized and analyzed, door lock power supply strategy control parameters are obtained, self-adaptive power supply management and control are conducted on the door lock, the technical problems that an existing self-powered intelligent door lock is unstable in electric quantity and difficult to accurately predict and effectively manage are solved, efficient utilization and intelligent scheduling of the electric quantity of the door lock are achieved, and the safety of the door lock is improved. And the autonomous power supply capability and the operation efficiency of the door lock system are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of door lock power supply control, and in particular to an adaptive power supply control method and device for a self-powered intelligent door lock. Background Art

[0002] As an important part of smart home, smart door locks have been widely used in daily life due to their convenience, safety and convenience. However, the contradiction between the high frequency use and low power consumption design of smart door locks has always been an important bottleneck in its development. Efficient power management is the key to ensure the stable operation of smart door locks, extend the service life and improve the user experience. Although self-powered smart door locks have made certain progress in technology, there are still some problems. For example, the power consumption of smart door locks in dynamic and static states is very different. It is difficult for existing power management technologies to take into account both light and heavy load efficiency at the same time, resulting in high energy consumption in certain working states; in actual use, the reliability and stability of emergency power supply still need to be further improved; and how to maintain stable self-power supply capabilities in different environments is also a challenge.

[0003] In summary, existing self-powered smart door locks have technical problems such as unstable power supply, difficulty in accurate prediction and effective management. Summary of the invention

[0004] The present application provides an adaptive power supply control method and device for a self-powered smart door lock, which is used to solve the technical problems of unstable power supply, difficulty in accurate prediction and effective management of existing self-powered smart door locks.

[0005] In view of the above problems, the present application provides an adaptive power supply control method and device for a self-powered smart door lock.

[0006] In a first aspect, the present application provides an adaptive power supply control method for a self-powered smart door lock, the method comprising: A self-powered smart door lock is obtained, wherein the self-powered smart door lock has a built-in handle-pressing power generation device and an energy storage element. When the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electric energy through a speed increasing device and stores the electric energy in the energy storage element. A power usage data set of the self-powered smart door lock is collected and obtained, and the power usage data set is classified, identified, trained and fused using a recurrent neural network structure to construct an adaptive power consumption prediction network. The current power state of the energy storage element and the working state of the door lock are monitored and obtained through a door lock control module, and the power consumption of the door lock working state in a preset time window is predicted based on the adaptive power consumption prediction network to obtain the predicted power consumption parameters of the door lock. A door lock power supply control strategy library is constructed, and matching analysis is performed with the door lock power supply control strategy library based on the predicted power consumption parameters of the door lock and the current power state to determine a target door lock power supply control strategy. The predicted power consumption parameters of the door lock are optimized and analyzed using an adaptive power supply scheduling algorithm and the target door lock power supply control strategy to obtain the door lock power supply strategy control parameters, and the self-powered smart door lock is adaptively powered on and controlled based on the door lock power supply strategy control parameters.

[0007] In a second aspect, the present application provides an adaptive power supply control device for a self-powered smart door lock, the device comprising: The device acquisition module is used to acquire the self-powered intelligent door lock. The self-powered intelligent door lock has a built-in handle-pressing power generation device and an energy storage element. When the self-powered intelligent door lock is operated by the user, the handle-pressing power generation device generates electric energy through a speed-increasing device and stores it in the energy storage element. The data acquisition module is used to acquire the power usage data set of the self-powered intelligent door lock, classify and identify the power usage data set and perform training fusion using a recurrent neural network structure to construct an adaptive power consumption prediction network. The state monitoring module is used to monitor and acquire the current power state of the energy storage element and the working state of the door lock through the door lock control module, and based on the power consumption self- The adaptive prediction network predicts the power consumption of the door lock working state in a preset time window and obtains the predicted power consumption parameters of the door lock; the matching analysis module is used to build a door lock power supply control strategy library, and based on the door lock power supply prediction consumption parameters and the current power state, the matching analysis is performed with the door lock power supply control strategy library to determine the target door lock power supply control strategy; the power supply management and control module is used to use the adaptive power supply scheduling algorithm and the target door lock power supply control strategy to optimize and analyze the door lock power supply prediction consumption parameters, obtain the door lock power supply strategy control parameters, and perform adaptive power supply management and control on the self-powered smart door lock based on the door lock power supply strategy control parameters.

[0008] One or more technical solutions provided in this application have at least the following technical effects or advantages: The present application provides an adaptive power supply control method for a self-powered smart door lock, which obtains a self-powered smart door lock, wherein the self-powered smart door lock has a built-in handle-pressing power generation device and an energy storage element. When the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electric energy through an acceleration device and stores the electric energy in the energy storage element; a power supply usage data set of the self-powered smart door lock is collected and obtained, and a recurrent neural network structure is used to classify, identify and train the power supply usage data set to construct an adaptive power consumption prediction network; the current power state of the energy storage element and the working state of the door lock are monitored and obtained by a door lock control module, and the power consumption of the door lock working state is predicted within a preset time window based on the adaptive power consumption prediction network. Obtain the predicted power consumption parameters of the door lock; construct a door lock power supply control strategy library, and based on the predicted power consumption parameters of the door lock and the current power status, perform matching analysis with the door lock power supply control strategy library to determine the target door lock power supply control strategy; utilize an adaptive power supply scheduling algorithm and the target door lock power supply control strategy to optimize and analyze the predicted power consumption parameters of the door lock, obtain the door lock power supply strategy control parameters, and perform adaptive power supply management and control on the self-powered smart door lock based on the door lock power supply strategy control parameters, thereby solving the technical problems of unstable power supply, difficulty in accurate prediction and effective management of existing self-powered smart door locks, achieving efficient utilization and intelligent scheduling of door lock power, and improving the technical effect of autonomous power supply capability and operation efficiency of door lock systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Figure 1 A flow chart of an adaptive power supply control method for a self-powered smart door lock is provided for this application.

[0010] Figure 2 A schematic diagram of the structure of an adaptive power supply control device for a self-powered smart door lock is provided for this application.

[0011] Explanation of the reference numerals: device acquisition module 11 , data acquisition module 12 , status monitoring module 13 , matching analysis module 14 , power supply control module 15 . DETAILED DESCRIPTION

[0012] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0013] Embodiment 1 like Figure 1 As shown, the present application provides an adaptive power supply control method for a self-powered smart door lock, the method comprising: Step S100: Obtain a self-powered smart door lock, which has a built-in handle-pressing power generation device and an energy storage element. When the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electrical energy through a speed increasing device and stores it in the energy storage element.

[0014] Specifically, the self-powered smart door lock is an innovative door lock system that no longer relies entirely on an external power source (such as a battery or a power grid), but instead uses the mechanical energy generated by the user's daily operations (such as opening and closing the door) to convert it into electrical energy to achieve self-powering. This design not only reduces the trouble and cost of replacing batteries, but also improves the environmental protection and sustainability of the door lock. The self-powered smart door lock has a built-in push handle power generation device and an energy storage element. Among them, the push handle refers to the part of the smart door lock that the user holds and pushes to open or close the door body; the power generation device is a device that can convert mechanical energy (such as rotation, vibration, pressure, etc.) into electrical energy. When the user performs the door opening or closing operation, he will naturally hold and push the push handle. In this process, the rotation or linear motion of the push handle is converted into mechanical energy. In order to make more efficient use of this mechanical energy, a speed-increasing device is also integrated inside the push handle power generation device. The speed-increasing device (such as a gear set, etc.) amplifies the mechanical motion of the push handle by increasing the rotation speed or changing the direction of motion, thereby improving the power generation efficiency. The power generation device then converts this amplified mechanical energy into electrical energy.

[0015] The energy storage element built into the self-powered smart door lock is a device that can store electrical energy and release it when needed, such as a battery, supercapacitor, etc. The electrical energy generated by the power generation device is directly transmitted to the energy storage element for storage. The energy storage element plays the role of an energy warehouse in the smart door lock. It can store electrical energy when the user does not operate the door lock, and release electrical energy when the door lock needs power (such as identity authentication, driving the motor to unlock, etc.). Common energy storage elements include rechargeable batteries (such as lithium batteries) and supercapacitors. Compared with batteries, supercapacitors have higher charge and discharge rates and longer cycle life, but lower energy density; while batteries have higher energy density, but the charge and discharge rate and cycle life may be limited. Therefore, in practical applications, self-powered smart door locks can select appropriate energy storage elements or use them in combination according to specific needs.

[0016] Through the above steps, the self-powered smart door lock realizes energy recovery and reuse of users' daily operations, which not only reduces dependence on external power supply, but also improves the intelligence level of the door lock and user experience.

[0017] Step S200: Collect and obtain a power usage data set of the self-powered smart door lock, use a recurrent neural network structure to classify and identify the power usage data set and perform training fusion to build a power consumption adaptive prediction network.

[0018] Optionally, design a data acquisition system that can monitor and record the power consumption of the self-powered smart door lock in real time, including data acquisition of the internal sensors of the door lock (such as current sensors, voltage sensors), and possible external trigger signals (such as door opening signals, door closing signals). Data acquisition should cover the normal working cycle of the door lock, including power changes in different time periods, frequency and type of user operations, etc. The collected data will be used as a power usage data set for subsequent analysis and modeling. The power usage data set refers to a collection of data such as power consumption, charging status, user operation mode, etc. of the self-powered smart door lock in different time periods and different usage scenarios.

[0019] Furthermore, a recurrent neural network structure is used to classify and identify the power usage data set and train and fuse it to build an adaptive prediction network for power consumption. Before using the data set for model training, data preprocessing is required, including steps such as data cleaning, feature extraction and standardization, so as to eliminate noise, reduce data redundancy, and improve the effect of model training. In the specific modeling process, in order to classify and identify the power usage data set and train and fuse it, a recurrent neural network (RNN) can be selected as the basic structure of the model. The recurrent neural network (RNN) is a special type of neural network that is good at processing sequence data. RNN introduces recurrent connections in the network so that the network can remember previous information, which is suitable for time series analysis and prediction tasks. First, according to the characteristics of the data set, the structure and parameters of the input layer, hidden layer and output layer of the RNN are determined. The input layer receives the preprocessed data sequence, the hidden layer captures the temporal dependency in the data through recurrent connections, and the output layer performs classification identification or prediction based on the output of the hidden layer.

[0020] In the RNN model, classification identification can be achieved by adding a softmax classifier in the output layer. The softmax classifier can convert the output of the RNN into a probability distribution to determine the category to which each sample belongs. However, in the power usage data set of the self-powered smart door lock, more attention may be paid to the prediction of power consumption rather than simple classification. Therefore, the training fusion process can be more reflected in the optimization of the RNN model parameters. Through optimization methods such as back propagation algorithm and gradient descent, the model parameters are continuously adjusted to minimize the prediction error.

[0021] Finally, based on the trained RNN model, an adaptive power consumption prediction network is constructed. The network can receive real-time power usage data as input and use the RNN's time series modeling capability to predict power consumption in the future. In order to achieve adaptive prediction, the network can also automatically adjust the prediction parameters and model structure according to historical data and the current environment (such as user operation mode, door lock status, etc.). In this way, the adaptive power consumption prediction network can provide more accurate and reliable power management support for self-powered smart door locks.

[0022] By building an adaptive power consumption prediction network, we can not only improve the intelligence level of door locks, but also provide users with more convenient and efficient power management services.

[0023] Step S300: The current power state of the energy storage element and the door lock working state are monitored and acquired through the door lock control module, and the power consumption of the door lock working state is predicted in a preset time window based on the power consumption adaptive prediction network to acquire the door lock power consumption prediction parameters.

[0024] For example, the door lock control module is the core component of the door lock system, which is responsible for controlling the unlocking and locking operations of the door lock, and is integrated with the monitoring function of the power status of the energy storage element (such as the battery) and the working status of the door lock. The door lock control module obtains the current power of the battery in real time through the built-in power monitoring circuit or sensor, such as voltage, current and other parameters, so as to determine the remaining power of the battery, and monitor the working status of the door lock at the same time. The working status of the door lock refers to the current open / closed state of the door lock, and whether the unlocking or locking operation is being performed, etc. These status information are crucial for power consumption prediction.

[0025] The power consumption adaptive prediction network can adaptively predict the power consumption in the future based on historical power consumption data, current power status, door lock working status and other information. During the prediction, based on the power consumption adaptive prediction network, input parameters such as the current power status, door lock working status and preset time window, perform power consumption prediction, and obtain the door lock power consumption prediction parameters. The preset time window refers to a time period set by the user or the system, which is used to predict the power consumption of the door lock within the time period. The length of the time window can be adjusted according to actual needs.

[0026] In practical applications, power consumption prediction can help users or systems plan door lock maintenance in advance to avoid failure or damage to door locks due to insufficient power. At the same time, by continuously optimizing the prediction model, the accuracy of the prediction can be improved, providing users with a more reliable and intelligent door lock experience.

[0027] Step S400: construct a door lock power supply control strategy library, and perform matching analysis with the door lock power supply control strategy library based on the door lock power consumption prediction parameter and the current power state to determine the target door lock power supply control strategy.

[0028] Furthermore, a door lock power supply control strategy library is constructed to cope with power supply management requirements under different power states and predicted consumption conditions. The strategy library contains a series of preset control strategies, each of which is designed for specific power conditions and prediction results, aiming to optimize the power usage of the door lock, extend the battery life, and ensure that the door lock can work normally at critical moments. Once the predicted power consumption parameters and current power state of the door lock are obtained, the system can start the matching analysis process, which will comprehensively consider the predicted power consumption parameters (such as the expected power consumption in the future), the current power state (such as the remaining power percentage) and possible external factors (such as user usage habits, ambient temperature, etc.), and then search for matching control strategies in the door lock power supply control strategy library. Through matching analysis, the control strategy that best matches the current predicted power consumption parameters and power state is found in the door lock power supply control strategy library, and it is determined as the target door lock power supply control strategy. This strategy will guide the power management of the door lock in the subsequent working process, including whether to start the energy-saving mode, when to remind the user to replace the battery, and whether to restrict the use of certain functions when the power is extremely low. For example, when the power is sufficient and the future power consumption is low, the control module reduces the discharge rate of the energy storage element to extend its service life; when the power is insufficient or the future power consumption is high, the control module prioritizes shutting down or reducing the power supply of non-essential functions and increasing the discharge rate of the energy storage element to meet the working requirements of the door lock.

[0029] By building a power supply control strategy library, combining the door lock power prediction with the current status, and intelligently matching and determining the optimal power supply strategy, the battery life can be effectively extended and the stable operation of the door lock can be ensured.

[0030] Step S500: Utilize the adaptive power supply scheduling algorithm and the target door lock power supply control strategy to optimize and analyze the predicted power consumption parameters of the door lock, obtain the door lock power supply strategy control parameters, and adaptively control the power supply of the self-powered smart door lock based on the door lock power supply strategy control parameters.

[0031] Specifically, after determining the target door lock power supply control strategy, the adaptive power supply scheduling algorithm can be used to further optimize the execution details of this strategy. The adaptive power supply scheduling algorithm will dynamically adjust the control parameters of the power supply strategy according to the real-time changes in the predicted power consumption parameters of the door lock to more accurately match the actual usage and power demand of the door lock. This algorithm can comprehensively consider the historical usage data of the door lock, the current power status, the predicted consumption trend, and external environmental factors, and find the optimal power supply management solution through intelligent analysis and calculation.

[0032] Under the framework of the adaptive power supply scheduling algorithm, the predicted power consumption parameters of the door lock are optimized and analyzed, aiming to find the best control parameter combination that can maximize the power utilization efficiency of the door lock, extend battery life, or meet other specific goals through detailed analysis and calculation of the predicted consumption parameters. After the optimization analysis, a set of optimized door lock power supply strategy control parameters can be obtained. These parameters will serve as the basis for subsequent adaptive power supply management and control, and guide how to manage power more reasonably during actual use of the door lock. These control parameters may include power thresholds, charging timing, working mode switching conditions, etc., which together constitute the door lock power supply strategy.

[0033] Finally, according to the obtained control parameters of the door lock power supply strategy, the self-powered smart door lock is adaptively controlled. This process will monitor the power status and predicted consumption of the door lock in real time, and automatically adjust the power supply behavior of the door lock according to the setting of the control parameters. For example, when the power is lower than a certain threshold, the user is automatically reminded to replace the battery, or the energy-saving mode is started in advance when the power is sufficient but the future consumption is predicted to be high. In this way, it can ensure that the door lock can maintain the best power status and working performance in different usage scenarios.

[0034] Furthermore, the step S200 of constructing the power consumption adaptive prediction network in the present application further includes: Step S210 - 1 : Obtain information on factors affecting door lock power supply, where the information on factors affecting door lock power supply includes power consumption, user operation parameters, environmental parameters, and door lock performance parameters.

[0035] Step S220 - 1 : classifying and identifying the power supply usage data set based on the door lock power supply influencing factor information to obtain a door lock power supply usage factor data set.

[0036] Step S230 - 1 : Sequentially divide the door lock power supply usage factor data set according to the preset time window to obtain a power supply factor sequence usage data set.

[0037] Step S240-1: Use a recurrent neural network structure to perform supervised training on the power supply factor sequence using a data set to obtain a power consumption factor prediction branch network set.

[0038] Step S250 - 1 : weightedly merging each branch network in the power consumption factor prediction branch network set to construct the power consumption adaptive prediction network.

[0039] Optionally, first of all, it is necessary to comprehensively collect information on various factors affecting the power consumption of door locks, including but not limited to power consumption data (such as historical power consumption records, including daily power consumption and power consumption in different time periods), user operation parameters (such as door opening frequency, operation time distribution), environmental parameters (such as temperature, etc.) and door lock performance parameters (such as battery aging, etc.). These data will serve as the basis for subsequent analysis and help build an accurate power consumption prediction model. After collecting the information on factors affecting the power supply of door locks, these data need to be preprocessed and classified to obtain a data set of factors for the use of power supply for door locks. Specifically, according to the different characteristics and sources of the data, they are divided into different categories or subsets for more targeted analysis and modeling. The classification and identification process involves steps such as data cleaning, conversion, and labeling to ensure the accuracy and consistency of the data.

[0040] Furthermore, in order to capture the trend and law of the power consumption of the door lock over time, it is necessary to divide the power usage factor data set into sequences according to the preset time window. The size of the time window can be determined according to actual needs and data characteristics, and should generally be able to cover the main change cycle of the power consumption of the door lock. Through sequence division, the data set can be converted into a series of ordered time series data, providing a basis for subsequent prediction using models such as recurrent neural networks.

[0041] After obtaining the power factor sequence usage dataset, the next step is to supervise the training of these datasets using deep learning models such as recurrent neural networks (RNNs). By dividing and training the datasets corresponding to the information of different door lock power influencing factors, a prediction branch network dedicated to the factor can be obtained to predict the power consumption under the influence of the factor. Recurrent neural networks are particularly suitable for processing time series data because they can capture the temporal dependencies and long-term memory effects in the data. During the training process, the model learns how to predict future power consumption based on the input time series data. By continuously adjusting the model parameters and optimizing the training process, a power consumption prediction model with high prediction accuracy can be obtained.

[0042] In order to improve the accuracy and robustness of power consumption prediction, multiple power consumption prediction models (i.e., branch networks) based on different features or data sets can be weighted and fused. The weighted fusion process involves determining the weight coefficients of each branch network, which reflect the importance or contribution of each branch network in the final prediction result. By reasonably setting the weight coefficients and performing weighted summation or weighted averaging of the prediction results of each branch network, a more comprehensive and accurate power consumption prediction result can be obtained.

[0043] Furthermore, the step S200 of constructing the power consumption adaptive prediction network in the present application further includes: Step S210 - 2 : performing multi-dimensional performance verification on each branch network in the power consumption factor prediction branch network set to obtain network prediction multi-dimensional performance parameters.

[0044] Step S220 - 2 : performing reliability evaluation on the power consumption factor prediction branch network set based on the network prediction multi-dimensional performance parameters, and determining a branch network reliability coefficient set.

[0045] Step S230-2: weight the door lock power supply influencing factor information, obtain a branch network weight factor set, and use the product of the branch network reliability coefficient set and the branch network weight factor set as the branch network decision parameter set.

[0046] Step S240 - 2 : performing weighted fusion on the power consumption factor prediction branch network set based on the branch network decision parameter set to obtain the power consumption adaptive prediction network.

[0047] Furthermore, after obtaining the set of power consumption factor prediction branch networks, in order to ensure the prediction accuracy and stability of each branch network, multi-dimensional performance verification is required. This step involves evaluating the performance of the branch network in different aspects, including but not limited to prediction accuracy, robustness, generalization ability, etc. The verification process may include testing each branch network using an independent test data set, comparing the difference between their prediction results and actual values, and calculating the corresponding performance indicators (such as mean square error, accuracy, etc.). Through multi-dimensional performance verification, the multi-dimensional performance parameters of each branch network can be obtained, which will serve as the basis for subsequent reliability evaluation.

[0048] Based on the results of multi-dimensional performance verification, the reliability of the power consumption factor prediction branch network set needs to be evaluated next. The purpose of reliability evaluation is to determine the stability and credibility of each branch network in the prediction process. By comprehensively considering the multi-dimensional performance parameters of each branch network (such as prediction accuracy, error range, stability index, etc.), the reliability coefficient of each branch network can be calculated. These coefficients reflect the reliability and importance of each branch network in predicting power consumption.

[0049] After determining the reliability coefficient of the branch network, it is also necessary to assign weights to the information of the factors affecting the power supply of the door lock. Since different influencing factors contribute differently to power consumption, they need to be assigned different weight factors. The larger the factor weight, the higher the performance reliability of the branch network, and the greater the decision-making degree in the final model. Weight assignment can be based on expert experience, historical data analysis, or other appropriate methods. Through weight assignment, a set of branch network weight factors can be obtained. Each element in the set corresponds to a branch network and reflects the relative importance of the branch network in the final prediction result.

[0050] In order to comprehensively consider the reliability of the branch network and the importance of the influencing factors, the branch network reliability coefficient set and the branch network weight factor set are element-by-element multiplied, and the product result is the branch network decision parameter set. Each element in the branch network decision parameter set reflects the reliability of the branch network and considers the importance of its corresponding influencing factors, providing a decision basis for subsequent weighted fusion.

[0051] Finally, the power consumption factor prediction branch network set is weighted and fused based on the branch network decision parameter set. The weighted fusion process is to perform weighted summation or weighted average operations on the prediction results of each branch network according to its corresponding decision parameters. In this way, the prediction results not only consider the prediction ability of each branch network, but also fully consider their reliability and the importance of influencing factors. Through weighted fusion, a more comprehensive, accurate and reliable power consumption adaptive prediction network can be obtained. The network can dynamically adjust the prediction model according to the real-time situation of the door lock power supply and the changes in influencing factors, and realize the adaptive prediction of power consumption.

[0052] Furthermore, the step S240-2 of obtaining the power consumption adaptive prediction network also includes: Step S241 - 2 : performing weighted fusion on the power consumption factor prediction branch network set based on the branch network decision parameter set to obtain an initial power consumption prediction network.

[0053] Step S242-2: Perform loss assessment on the initial power consumption prediction network, determine network performance loss parameters, perform network parameter gradient calculation on the initial power consumption prediction network based on the network performance loss parameters, and obtain network parameter gradient information.

[0054] Step S243-2: Analyze the optimization direction of the network parameter gradient information, determine the network parameter optimization strategy, and iteratively update the initial power consumption prediction network based on the network parameter optimization strategy using a back propagation algorithm to obtain network optimization parameters.

[0055] Step S244 - 2 : updating the configuration of the initial power consumption prediction network based on the network optimization parameters to obtain the power consumption adaptive prediction network.

[0056] Exemplarily, the power consumption factor prediction branch network set is weighted and fused according to the calculated branch network decision parameter set. Specifically, the prediction results of each branch network are weighted summed or weighted averaged according to their corresponding decision parameters, so as to obtain an initial power consumption prediction network that integrates the prediction capabilities of multiple branch networks. This network initially has the ability to predict power consumption according to the factors affecting the power supply of the door lock.

[0057] Next, it is necessary to perform a loss assessment on the initial power consumption prediction network. The purpose of loss assessment is to quantify the difference between the network prediction result and the actual value, that is, the prediction error of the network. Commonly used loss functions include mean square error (MSE), mean absolute error (MAE), etc. By calculating the value of the loss function, the network performance loss parameter can be obtained. This parameter reflects the prediction accuracy and performance of the network in the current state. Further, the network parameter gradient is calculated. The network parameter gradient refers to the partial derivative of the loss function with respect to the network parameters, which indicates the trend and direction of the loss function as the network parameters change. The gradient information of the network parameters can be efficiently calculated through the back propagation algorithm, and this gradient information will be used to guide the optimization and adjustment of the network parameters.

[0058] After obtaining the network parameter gradient information, it is necessary to analyze it to determine the optimization direction and optimization strategy of the network parameters. The choice of optimization strategy depends on the specific optimization algorithm, such as stochastic gradient descent (SGD), Adam, etc. These optimization algorithms will calculate the update amount of the network parameters based on the gradient information and guide the network parameters to adjust in the direction of reducing the loss function. Through the analysis and determination of the optimization strategy, guidance can be provided for the subsequent iterative update process. After determining the network parameter optimization strategy, the back propagation algorithm is used to iteratively update the initial power consumption prediction network. The iterative update process is the process of reducing the loss function value by continuously adjusting the network parameters. In each round of iteration, new network parameter values ​​are calculated based on the current network parameters and gradient information, and these new parameter values ​​are applied to the network. Through multiple iterative updates, the network parameters can be gradually optimized to improve the prediction accuracy and performance of the network. For example, according to the optimization strategy, the model learning rate can be added to the update process to control the update step size of the parameters, and the model parameter update steps are repeatedly performed to gradually optimize the model parameters. The value of the loss function is reduced by iteratively adjusting the model parameters to minimize the value of the loss function.

[0059] Finally, based on the network optimization parameters obtained during the iterative update process, the initial power consumption prediction network is configured and updated. This step is to apply the optimized network parameters to the network and replace the original parameter values. The network after the configuration update is the final power consumption adaptive prediction network. This network can dynamically adjust the prediction model according to the real-time situation of the door lock power supply and changes in influencing factors to achieve adaptive prediction of power consumption. At the same time, due to multiple iterative optimization and performance evaluation, the network has high prediction accuracy and robustness.

[0060] Furthermore, the step S400 of determining the target door lock power supply control strategy also includes: Step S410: obtaining power supply strategy classification rules, wherein the power supply strategy classification rules include power consumption classification threshold rules and power status classification threshold rules.

[0061] Step S420: encoding the category parameters based on the power consumption classification threshold rule and the power status classification threshold rule, and determining the strategy threshold calibration encoding parameters.

[0062] Step S430: dividing and calibrating the door lock power supply control strategy library based on the strategy threshold calibration coding parameter to obtain a door lock power supply calibration strategy library.

[0063] Step S440: matching the predicted power consumption parameters of the door lock and the current power status with the door lock power supply calibration strategy library to determine the target door lock power supply control strategy.

[0064] Furthermore, in order to formulate an effective door lock power supply control strategy, it is necessary to obtain the pre-set power supply strategy division rules. These rules are formulated based on the operating characteristics and power management requirements of the door lock system, and mainly include power consumption classification threshold rules and power state classification threshold rules, that is, the rules characterize the power consumption threshold division basis and power state threshold division basis corresponding to different types of power supply strategies. Among them, the power consumption classification threshold rule defines the classification standards under different power consumption levels, such as low consumption, medium consumption, high consumption, etc., and the corresponding threshold limits; while the power state classification threshold rule defines the classification standards for the current power of the door lock, such as sufficient, medium, low power, etc., which also include clear threshold limits. These rules provide the basis for subsequent category parameter encoding and strategy library division, so that in actual control, the power supply of non-essential functions can be preferentially turned off or reduced, thereby increasing the discharge rate of the energy storage element.

[0065] Next, the category parameters are encoded according to the obtained power consumption classification threshold rules and power state classification threshold rules. The purpose of encoding is to convert abstract classification standards and threshold boundaries into computer-recognizable codes or identifiers. For power consumption classification, a unique code can be assigned to each consumption level; for power state classification, a code is also assigned to each state level. Through encoding, data classification, comparison and matching can be easily achieved. In addition, it is also necessary to determine the strategy threshold calibration coding parameters, that is, the relationship between these codes and specific control strategies, to provide a basis for subsequent strategy library division and parameter matching.

[0066] After completing the category parameter encoding, the door lock power supply control strategy library is divided and calibrated according to the strategy threshold calibration coding parameters. The strategy library is a collection of multiple power supply control strategies, which formulate corresponding control measures for different power consumption and power status conditions. Through division and calibration, the strategies in the strategy library can be organized in order according to the power consumption classification and power status classification to form a door lock power supply calibration strategy library. This calibration strategy library provides a convenient query and matching mechanism for the subsequent target strategy determination.

[0067] Finally, the predicted power consumption parameters of the door lock (i.e., the prediction results obtained based on the power consumption adaptive prediction network) and the current power status (obtained through real-time monitoring of the door lock system) are used as inputs to match the parameters with the door lock power supply calibration strategy library. The matching process is achieved by comparing whether the predicted consumption parameters and the current power status match the threshold calibration encoding parameters of each strategy in the calibration strategy library. When a matching strategy is found, the target door lock power supply control strategy is determined, which will serve as the guiding principle for the subsequent power supply control of the door lock system to ensure that the door lock reaches the optimal state in terms of power management.

[0068] Furthermore, the obtaining of the door lock power supply strategy control parameter, step S500 of the present application further includes: Step S510: Perform historical data mining based on the target door lock power supply control strategy to construct a door lock power supply strategy control parameter space.

[0069] Step S520: dividing the door lock power supply strategy control parameter space into control thresholds according to the door lock power consumption prediction parameters, and determining a control parameter selection memory library.

[0070] Step S530: construct a door lock power supply effect fitness function, and perform parameter search and evaluation in the control parameter selection memory library through the door lock power supply effect fitness function to obtain multiple control parameter fitnesses.

[0071] Step S540: Utilize an adaptive power supply scheduling algorithm to perform parameter optimization analysis on the fitness of the multiple control parameters to obtain the door lock power supply strategy control parameters.

[0072] Optionally, in order to obtain effective door lock power supply strategy control parameters, historical data mining is required based on the determined target door lock power supply control strategy. This step involves collecting and analyzing data generated during the past door lock power supply process, including power consumption records, power supply strategy execution effects, environmental condition changes and other information. Through data mining techniques, such as cluster analysis and association rule mining, characteristic patterns and laws related to the target power supply control strategy can be extracted from historical data. Then, a door lock power supply strategy control parameter space is constructed based on the results of historical data mining. This space is a multi-dimensional parameter set that includes various control parameters that affect the door lock power supply effect, such as power supply voltage, current, frequency, power supply duration, etc., where each parameter has a certain range of values, forming a dimension of the parameter space. By constructing the parameter space, the impact of different control parameter combinations on the door lock power supply effect can be systematically represented and evaluated.

[0073] After the control parameter space is constructed, the control threshold is divided into the space using the predicted power consumption parameters of the door lock. The purpose of this step is to divide the control parameter space into different regions or subsets according to the predicted power consumption, and each region corresponds to a set of specific control parameter value ranges. In this way, when the power consumption of the door lock falls into a certain interval, the control parameters can be selected from the corresponding parameter area to achieve more accurate power supply control. Furthermore, through the control threshold division, one or more control parameter selection memory banks can be obtained. These memory banks are actually subspaces or parameter sets divided in the control parameter space according to different power consumption conditions. Each memory bank contains a set of verified and optimized control parameters that show good power supply effects under similar power consumption conditions. The establishment of the memory bank provides an efficient index and reference for subsequent parameter search and evaluation.

[0074] Next, in order to evaluate the power supply effect of different control parameter combinations, a door lock power supply effect fitness function can be constructed. This function can be a mathematical model used to quantify the relationship between control parameters and power supply effect. The fitness function can be constructed based on multiple evaluation indicators, such as power supply efficiency, energy consumption cost, door lock service life, etc. By inputting different control parameter combinations into the fitness function, a numerical score reflecting the power supply effect can be obtained for subsequent comparison and optimization. After determining the control parameter selection memory library and the door lock power supply effect fitness function, the parameter search evaluation is then performed in the memory library. By traversing the control parameter combinations in the memory library, each combination is input into the fitness function for calculation to obtain the corresponding power supply effect score. By comparing the scoring results of different combinations, the control parameter combinations with excellent performance can be preliminarily screened out.

[0075] Finally, the adaptive power supply scheduling algorithm is used to conduct in-depth parameter optimization analysis on the fitness of multiple control parameters. The adaptive power supply scheduling algorithm is an optimization algorithm that can automatically adjust the control parameters based on real-time data and feedback. It can further explore the subtle differences in the parameter space based on the excellent control parameter combinations that have been screened out, and find more accurate and efficient power supply control parameters. Through multiple iterations and optimizations, the optimal door lock power supply strategy control parameters are finally obtained, providing strong support for the stable and efficient operation of the door lock system.

[0076] Furthermore, the obtaining of the door lock power supply strategy control parameter, step S540 of the present application further includes: Step S541: Initializing particle swarm parameters according to the adaptive power supply scheduling algorithm, the particle swarm parameters including particle position and particle velocity.

[0077] Step S542: performing gradient ascent calculation and optimization direction analysis on the fitness of the multiple control parameters to determine the optimization direction of the control parameters.

[0078] Step S543: the particle swarm parameters are evaluated and iteratively updated based on the control parameter optimization direction until the preset iteration conditions are met, and the particle with the largest fitness is optimized and determined as the door lock power supply strategy control parameter.

[0079] Specifically, in the process of using the adaptive power supply scheduling algorithm to find the control parameters of the door lock power supply strategy, it is necessary to initialize the particle swarm parameters first. The particle swarm optimization algorithm is an optimization method based on swarm intelligence. It simulates the foraging behavior of bird flocks to find the optimal solution. In this algorithm, each particle represents a potential solution, that is, a set of control parameter values. During initialization, an initial position and initial velocity need to be randomly assigned to each particle. The initial position is randomly selected in the control parameter space, indicating the initial control parameter combination represented by the particle; the initial velocity determines the direction and speed of the particle's movement in the search space.

[0080] Next, using the previously constructed door lock power supply effect fitness function, the fitness of multiple control parameters is calculated by gradient ascent and the optimization direction is analyzed. Gradient ascent is an optimization algorithm used to find the maximum value of a function. Here, the optimization direction of each control parameter can be determined by calculating the gradient of the fitness function with respect to each control parameter. The gradient points to the direction in which the function value increases the fastest, so adjusting the control parameters along the direction of the gradient is expected to improve the fitness of the power supply effect. By comprehensively analyzing the gradients of multiple control parameters, the optimization direction of the entire control parameter combination can be determined.

[0081] After determining the optimization direction of the control parameters, the particle swarm parameters are evaluated and iteratively updated based on this direction. Each particle updates its position and velocity based on its own experience and the experience of the entire particle swarm (i.e., the position and velocity information of other particles). Specifically, the particle will move to the currently discovered optimal position (i.e., the position of the particle with the highest fitness) and its own historical optimal position, while maintaining a certain degree of randomness to avoid falling into the local optimum. By continuously iteratively updating the particle swarm parameters, the particles will gradually gather in areas with higher fitness and eventually approach the global optimal solution.

[0082] The iterative update process will continue until the preset iteration conditions are met. These conditions can be that the maximum number of iterations is reached, the fitness change is less than the preset threshold, or the particle swarm is sufficiently concentrated. When one of these conditions is met, the iteration process ends. At this time, the control parameter combination represented by the particle with the largest fitness is considered to be the optimal control parameter of the door lock power supply strategy. This set of parameters performs best in terms of power supply effect, can meet the operating requirements of the door lock system to the greatest extent, and improve power supply efficiency and stability.

[0083] Through the technical solutions of the above-mentioned embodiments, the present application provides an adaptive power supply control method for a self-powered smart door lock, which solves the technical problems of unstable power supply, difficulty in accurate prediction and effective management of existing self-powered smart door locks, achieves efficient utilization and intelligent scheduling of power in door locks, and improves the technical effect of autonomous power supply capability and operating efficiency of the door lock system.

[0084] Embodiment 2 Based on the same inventive concept as the adaptive power supply control method of a self-powered smart door lock in the aforementioned embodiment, Figure 2 As shown, the present application provides an adaptive power supply control device for a self-powered intelligent door lock, the device comprising: The device acquisition module 11 is used to obtain a self-powered smart door lock, which has a built-in handle-pressing power generation device and an energy storage element. When the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electrical energy through a speed increasing device and stores it in the energy storage element.

[0085] The data acquisition module 12 is used to collect and obtain the power usage data set of the self-powered intelligent door lock, classify and identify the power usage data set and perform training fusion using a recurrent neural network structure to construct an adaptive power consumption prediction network.

[0086] The state monitoring module 13 is used to monitor and obtain the current power state of the energy storage element and the working state of the door lock through the door lock control module, predict the power consumption of the door lock working state in a preset time window based on the power consumption adaptive prediction network, and obtain the door lock power consumption prediction parameters.

[0087] The matching analysis module 14 is used to construct a door lock power supply control strategy library, and based on the door lock power consumption prediction parameter and the current power state, performs matching analysis with the door lock power supply control strategy library to determine a target door lock power supply control strategy.

[0088] The power supply control module 15 is used to use the adaptive power supply scheduling algorithm and the target door lock power supply control strategy to optimize and analyze the predicted power consumption parameters of the door lock, obtain the door lock power supply strategy control parameters, and adaptively control the power supply of the self-powered smart door lock based on the door lock power supply strategy control parameters.

[0089] Furthermore, the data acquisition module 12 is further configured to perform the following steps: Obtain information about factors affecting door lock power supply, where the information about factors affecting door lock power supply includes power consumption, user operation parameters, environmental parameters, and door lock performance parameters.

[0090] The power supply usage data set is classified and identified based on the door lock power supply influencing factor information to obtain a door lock power supply usage factor data set.

[0091] The door lock power supply usage factor data set is divided into sequences according to the preset time window to obtain a power supply factor sequence usage data set.

[0092] A recurrent neural network structure is used to perform supervised training on the power supply factor sequence using a data set to obtain a branch network set for predicting power consumption factors.

[0093] The branch networks in the power consumption factor prediction branch network set are weighted and fused to construct the power consumption adaptive prediction network.

[0094] Furthermore, the data acquisition module 12 is further configured to perform the following steps: A multi-dimensional performance verification is performed on each branch network in the power consumption factor prediction branch network set to obtain a network prediction multi-dimensional performance parameter.

[0095] Based on the network prediction multi-dimensional performance parameters, the reliability of the power consumption factor prediction branch network set is evaluated to determine the branch network reliability coefficient set.

[0096] The door lock power supply influencing factor information is weighted, a branch network weight factor set is obtained, and the product of the branch network reliability coefficient set and the branch network weight factor set is used as the branch network decision parameter set.

[0097] The power consumption factor prediction branch network set is weightedly fused based on the branch network decision parameter set to obtain the power consumption adaptive prediction network.

[0098] Furthermore, the data acquisition module 12 is further configured to perform the following steps: The power consumption factor prediction branch network set is weightedly fused based on the branch network decision parameter set to obtain an initial power consumption prediction network.

[0099] Perform a loss assessment on the initial power consumption prediction network, determine a network performance loss parameter, perform a network parameter gradient calculation on the initial power consumption prediction network based on the network performance loss parameter, and obtain network parameter gradient information.

[0100] The optimization direction of the network parameter gradient information is analyzed to determine the network parameter optimization strategy, and the initial power consumption prediction network is iteratively updated based on the network parameter optimization strategy using a back propagation algorithm to obtain network optimization parameters.

[0101] The initial power consumption prediction network is configured and updated based on the network optimization parameters to obtain the power consumption adaptive prediction network.

[0102] Furthermore, the matching analysis module 14 is further configured to perform the following steps: A power supply strategy classification rule is obtained, where the power supply strategy classification rule includes a power consumption classification threshold rule and a power state classification threshold rule.

[0103] Based on the power consumption classification threshold rule and the power state classification threshold rule, category parameter encoding is performed to determine the strategy threshold calibration encoding parameter.

[0104] The door lock power supply control strategy library is divided and calibrated based on the strategy threshold calibration coding parameter to obtain a door lock power supply calibration strategy library.

[0105] The predicted power consumption parameters of the door lock and the current power status are matched with the door lock power supply calibration strategy library to determine the target door lock power supply control strategy.

[0106] Furthermore, the power supply control module 15 is further configured to perform the following steps: Historical data mining is performed based on the target door lock power supply control strategy to construct a door lock power supply strategy control parameter space.

[0107] The control parameter space of the door lock power supply strategy is divided into control thresholds according to the predicted power consumption parameters of the door lock, and a control parameter selection memory library is determined.

[0108] A door lock power supply effect fitness function is constructed, and parameter search and evaluation is performed in the control parameter selection memory library through the door lock power supply effect fitness function to obtain multiple control parameter fitnesses.

[0109] An adaptive power supply scheduling algorithm is used to perform parameter optimization analysis on the fitness of the multiple control parameters to obtain the control parameters of the door lock power supply strategy.

[0110] Furthermore, the power supply control module 15 is further configured to perform the following steps: Initialize particle swarm parameters according to the adaptive power supply scheduling algorithm, and the particle swarm parameters include particle positions and particle speeds.

[0111] The fitness of the multiple control parameters is subjected to gradient ascent calculation and optimization direction analysis to determine the optimization direction of the control parameters.

[0112] The particle swarm parameters are evaluated and iteratively updated based on the control parameter optimization direction until a preset iteration condition is met, and the particle with the largest fitness is determined by optimization as the door lock power supply strategy control parameter.

[0113] Through the above-mentioned detailed description of the adaptive power supply control method of a self-powered intelligent door lock in this specification, those skilled in the art can clearly know the adaptive power supply control device of a self-powered intelligent door lock in this embodiment. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.

[0114] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. An adaptive power supply control method for a self-powered intelligent door lock, characterized in that: The method comprises: Obtain a self-powered smart door lock, wherein the self-powered smart door lock has a built-in handle-pressing power generation device and an energy storage element, and when the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electric energy through a speed increasing device and stores the electric energy in the energy storage element; Collect and obtain a power usage data set of the self-powered smart door lock, classify and identify the power usage data set and perform training fusion using a recurrent neural network structure to build an adaptive power consumption prediction network; The current power state of the energy storage element and the working state of the door lock are monitored and acquired through the door lock control module, and the power consumption of the door lock working state is predicted in a preset time window based on the power consumption adaptive prediction network to acquire the power consumption prediction parameter of the door lock; Constructing a door lock power supply control strategy library, and performing matching analysis with the door lock power supply control strategy library based on the door lock power consumption prediction parameter and the current power state, to determine a target door lock power supply control strategy; The adaptive power supply scheduling algorithm and the target door lock power supply control strategy are used to optimize and analyze the predicted power consumption parameters of the door lock, obtain the door lock power supply strategy control parameters, and adaptively control the power supply of the self-powered smart door lock based on the door lock power supply strategy control parameters.

2. The method for adaptive power supply control of a self-powered intelligent door lock according to claim 1, characterized in that: The method of constructing an adaptive power consumption prediction network includes: Obtaining information about factors affecting door lock power supply, wherein the information about factors affecting door lock power supply includes power consumption, user operation parameters, environmental parameters, and door lock performance parameters; Classify and identify the power usage data set based on the door lock power supply influencing factor information to obtain a door lock power supply usage factor data set; Sequentially dividing the door lock power supply usage factor data set according to the preset time window to obtain a power supply factor sequence usage data set; Use a recurrent neural network structure to perform supervised training on the power supply factor sequence using a data set to obtain a branch network set for power consumption factor prediction; The branch networks in the power consumption factor prediction branch network set are weighted and fused to construct the power consumption adaptive prediction network.

3. The adaptive power supply control method of a self-powered intelligent door lock according to claim 2, characterized in that: The step of constructing the power consumption adaptive prediction network includes: Performing multi-dimensional performance verification on each branch network in the power consumption factor prediction branch network set to obtain network prediction multi-dimensional performance parameters; Based on the network prediction multi-dimensional performance parameters, reliability evaluation is performed on the power consumption factor prediction branch network set to determine a branch network reliability coefficient set; Performing weight allocation on the door lock power supply influencing factor information, obtaining a branch network weight factor set, and taking the product of the branch network reliability coefficient set and the branch network weight factor set as the branch network decision parameter set; The power consumption factor prediction branch network set is weightedly fused based on the branch network decision parameter set to obtain the power consumption adaptive prediction network.

4. The method for adaptive power supply control of a self-powered intelligent door lock according to claim 3, characterized in that: The step of obtaining the power consumption adaptive prediction network includes: Based on the branch network decision parameter set, the power consumption factor prediction branch network set is weightedly integrated to obtain an initial power consumption prediction network; Performing a loss assessment on the initial power consumption prediction network, determining a network performance loss parameter, and performing a network parameter gradient calculation on the initial power consumption prediction network based on the network performance loss parameter to obtain network parameter gradient information; Analyze the optimization direction of the network parameter gradient information, determine the network parameter optimization strategy, and iteratively update the initial power consumption prediction network based on the network parameter optimization strategy using a back propagation algorithm to obtain network optimization parameters; The initial power consumption prediction network is configured and updated based on the network optimization parameters to obtain the power consumption adaptive prediction network.

5. The method for adaptive power supply control of a self-powered intelligent door lock according to claim 1, characterized in that: Determining the target door lock power supply control strategy includes: Acquire power supply strategy classification rules, wherein the power supply strategy classification rules include power consumption classification threshold rules and power state classification threshold rules; Based on the power consumption classification threshold rule and the power state classification threshold rule, category parameter encoding is performed to determine the strategy threshold calibration encoding parameter; Based on the strategy threshold calibration coding parameter, the door lock power supply control strategy library is divided and calibrated to obtain a door lock power supply calibration strategy library; The predicted power consumption parameters of the door lock and the current power status are matched with the door lock power supply calibration strategy library to determine the target door lock power supply control strategy.

6. The method for adaptive power supply control of a self-powered intelligent door lock according to claim 1, characterized in that: The obtaining of the door lock power supply strategy control parameters includes: Perform historical data mining based on the target door lock power supply control strategy to construct a door lock power supply strategy control parameter space; The control parameter space of the door lock power supply strategy is divided into control thresholds by the predicted power consumption parameter of the door lock, and a control parameter selection memory library is determined; Constructing a door lock power supply effect fitness function, and performing parameter search and evaluation in the control parameter selection memory library through the door lock power supply effect fitness function to obtain multiple control parameter fitnesses; An adaptive power supply scheduling algorithm is used to perform parameter optimization analysis on the fitness of the multiple control parameters to obtain the control parameters of the door lock power supply strategy.

7. The method for adaptive power supply control of a self-powered intelligent door lock according to claim 6, characterized in that: The obtaining of the door lock power supply strategy control parameter includes: Initialize particle swarm parameters according to the adaptive power supply scheduling algorithm, wherein the particle swarm parameters include particle position and particle velocity; Performing gradient ascent calculation and optimization direction analysis on the fitness of the multiple control parameters to determine the optimization direction of the control parameters; The particle swarm parameters are evaluated and iteratively updated based on the control parameter optimization direction until a preset iteration condition is met, and the particle with the largest fitness is determined by optimization as the door lock power supply strategy control parameter.

8. An adaptive power supply control device for a self-powered intelligent door lock, characterized in that: A method for implementing an adaptive power supply control method for a self-powered intelligent door lock according to any one of claims 1 to 7, the device comprising: A device acquisition module, used to acquire a self-powered smart door lock, wherein the self-powered smart door lock has a built-in handle-pressing power generation device and an energy storage element. When the self-powered smart door lock is operated by a user, the handle-pressing power generation device generates electric energy through a speed increasing device and stores the electric energy in the energy storage element. A data acquisition module is used to acquire a power usage data set of the self-powered intelligent door lock, classify and identify the power usage data set and perform training fusion using a recurrent neural network structure to construct an adaptive prediction network for power consumption; A state monitoring module, used to monitor and obtain the current power state of the energy storage element and the working state of the door lock through the door lock control module, and to predict the power consumption of the door lock working state in a preset time window based on the power consumption adaptive prediction network, and to obtain the predicted power consumption parameters of the door lock; A matching analysis module, used to construct a door lock power supply control strategy library, and to perform matching analysis with the door lock power supply control strategy library based on the door lock power consumption prediction parameter and the current power state, so as to determine a target door lock power supply control strategy; The power supply management and control module is used to use the adaptive power supply scheduling algorithm and the target door lock power supply control strategy to optimize and analyze the predicted power consumption parameters of the door lock, obtain the door lock power supply strategy control parameters, and adaptively manage the power supply of the self-powered smart door lock based on the door lock power supply strategy control parameters.

Images