Intelligent power consumption management system for Internet of Things equipment

Abstract

The invention discloses an intelligent power consumption management system for Internet of Things equipment, which relates to the technical field of intelligent power consumption management systems and comprises a behavior acquisition module, a dynamic adjustment module, an energy consumption optimization module, an energy prediction module, an energy supply switching module and a system feedback module. The behavior acquisition module is used for monitoring the running state of equipment to obtain equipment activity feature data; the dynamic adjusting module is used for processing the equipment activity characteristic data by adopting a self-adaptive learning algorithm to obtain a dynamic dormancy threshold parameter; the energy consumption optimization module is used for applying the dynamic dormancy threshold parameter by adopting a multi-objective optimization method to obtain a power consumption distribution scheme of each functional module; the energy prediction module is used for analyzing the power consumption distribution scheme based on a neural network model in combination with environmental parameters to obtain an energy supply prediction result; and the energy supply switching module is used for responding to the energy supply prediction result by adopting a priority strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent power consumption management systems, and in particular to an intelligent power consumption management system for Internet of Things devices. Background Art

[0002] An intelligent power management system is a solution designed to optimize energy usage in IoT devices through intelligent means. It integrates multiple advanced technologies to effectively manage and control device power consumption. Therefore, leveraging advanced technologies to enhance the intelligence and security of intelligent power management systems has become a pressing issue.

[0003] In the field of intelligent power consumption management system technology, traditional power consumption management systems cannot simultaneously meet the requirements of fast response and low energy consumption. In scenarios that require rapid response, energy efficiency will be sacrificed. Conversely, the pursuit of energy saving will affect the system's immediate response capabilities. At the same time, most existing power consumption management systems use fixed thresholds and parameter settings, which are difficult to adapt to all changes in working modes or application scenarios, thereby affecting the overall system performance and user experience. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides an intelligent power consumption management system for IoT devices to solve the problem that most existing power consumption management systems use fixed thresholds and parameter settings, which are difficult to adapt to all changes in working modes or application scenarios, thereby affecting the overall performance of the system and user experience.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0007] In a first aspect, the present invention provides an intelligent power consumption management system for IoT devices, comprising:

[0008] Behavior collection module, dynamic adjustment module, energy consumption optimization module, energy prediction module, energy supply switching module and system feedback module;

[0009] The behavior collection module is used to monitor the operating status of the device and obtain device activity feature data;

[0010] The dynamic adjustment module is used to process the device activity feature data using an adaptive learning algorithm to obtain a dynamic sleep threshold parameter;

[0011] The energy consumption optimization module is used to apply a multi-objective optimization method to the dynamic sleep threshold parameters to obtain a power consumption allocation plan for each functional module;

[0012] The energy prediction module is used to analyze the power consumption allocation scheme in combination with environmental parameters based on a neural network model to obtain an energy supply prediction result;

[0013] The energy supply switching module is used to respond to the energy supply prediction result using a priority strategy to obtain an optimal energy supply mode instruction;

[0014] The system feedback module is used to evaluate the execution effect of the optimal energy supply mode instruction using a reinforcement learning method to obtain parameter optimization suggestions.

[0015] As a preferred solution of the intelligent power consumption management system for IoT devices of the present invention, wherein: the device operating status is monitored to obtain device activity characteristic data, the specific steps are as follows:

[0016] Set up a timed interrupt sampling mechanism to collect equipment operating status data at fixed time intervals;

[0017] The equipment operation status data includes the sending / receiving status flag of the communication module, the real-time load rate of the central processing unit and the data acquisition flag of the sensor module;

[0018] Calculate activity feature values ​​based on the collected equipment operating status data within a preset sliding time window;

[0019] The calculated activity characteristic values ​​are stored in a ring buffer, and the mean and standard deviation of the historical activity characteristic values ​​in the buffer are calculated;

[0020] When the buffer is full, device activity feature data is output, where the device activity feature data includes a current activity feature value, a historical activity feature mean, and a historical activity feature standard deviation.

[0021] As a preferred solution of the intelligent power consumption management system for IoT devices described in the present invention, wherein: the adaptive learning algorithm is used to process the device activity feature data to obtain the dynamic sleep threshold parameter, and the specific steps are as follows:

[0022] Calculating a dynamic sleep threshold parameter based on the received device activity characteristic data;

[0023] The calculated dynamic sleep threshold parameter T sleep Make boundary condition judgments and enforce them when specific conditions are met;

[0024] When A t >μ A +2σ A When T sleep Forced to 0.5 seconds;

[0025] When A t <μA -2σ A When T sleep Forced to 10 seconds;

[0026] The finalized dynamic sleep threshold parameter is output to the energy consumption optimization module.

[0027] As a preferred solution of the intelligent power consumption management system for IoT devices described in the present invention, the multi-objective optimization method is applied to the dynamic sleep threshold parameters to obtain the power consumption allocation scheme for each functional module. The specific steps are as follows:

[0028] Define a multi-objective function and calculate the optimal power consumption allocation solution based on the constraints configured in the multi-objective optimization function;

[0029] The linear programming algorithm is used to solve the above optimization problem to obtain the power consumption distribution value of each functional module, which includes the power consumption distribution value of the communication module P comm_opt , calculate the module allocation power consumption P comp_opt And the sensor module allocates power consumption P sense_opt .

[0030] As a preferred solution of the intelligent power consumption management system for IoT devices described in the present invention, the power consumption allocation scheme is analyzed in combination with environmental parameters based on a neural network model to obtain an energy supply prediction result. The specific steps are as follows:

[0031] The energy prediction module collects current environmental parameters, which include light intensity L, ambient temperature T, and vibration intensity V;

[0032] Input the received power consumption allocation plan and environmental parameters into the pre-trained LSTM neural network model;

[0033] The structure of the LSTM neural network model includes an input layer, a hidden layer and an output layer;

[0034] The input layer contains 5 neurons, corresponding to P comm_opt 、P comp_opt 、P sense_opt , light intensity and ambient temperature;

[0035] The hidden layer consists of two LSTM layers, each containing 32 neurons;

[0036] The output layer contains 3 neurons, which output the following prediction results respectively:

[0037] Solar energy input forecast value E in the next 1 hour solar ;

[0038] Vibration energy input prediction value E in the next 1 hourvib ;

[0039] Predicted value of total environmental energy input E total ;

[0040] The energy supply prediction result E is calculated by the LSTM neural network model total ;

[0041] The final calculated total environmental energy input prediction result E total The output is sent to the energy supply switching module for subsequent energy scheduling and power supply mode decision-making.

[0042] As a preferred solution of the intelligent power consumption management system for IoT devices described in the present invention, the LSTM neural network model is modeled and optimized through the following training process:

[0043] Collecting historical data samples, wherein the historical data samples include historical power consumption allocation plans, environmental parameters within a corresponding time period, and actually measured energy input data;

[0044] Normalize the collected data;

[0045] The model is trained using the Adam optimizer, and the loss function is defined as the sum of the mean square error (MSE) and the L2 regularization term.

[0046] The energy supply prediction cycle of the energy prediction module is consistent with the update cycle of the power consumption allocation scheme generated by the energy consumption optimization module, so as to ensure that the system achieves real-time matching between energy prediction and power consumption allocation during dynamic adjustment.

[0047] As a preferred solution of the intelligent power consumption management system for IoT devices of the present invention, wherein: the priority strategy is used to respond to the energy supply prediction result to obtain the optimal energy supply mode instruction, and the specific steps are as follows:

[0048] Determine the energy supply mode based on the priority strategy;

[0049] The priority strategy is specifically as follows:

[0050] When E total ≥E threshold_high When the system is in the state of emergency, the environment energy supply system is used first, and the excess energy is stored in the energy storage unit. At the same time, the energy supply mode instruction is output as the environment energy priority mode;

[0051] When E threshold_low ≤E total <E threshold_high When the power supply mode is combined with ambient energy and energy storage, the output power supply mode instruction is a hybrid power supply mode;

[0052] When Etotal <E threshold_low And S≥S critical When the power supply is in the energy storage mode, the energy storage priority mode is used first.

[0053] When E total <E threshold_low And S critical When the low power consumption operation mechanism is started, the output power supply mode instruction is the emergency energy saving mode;

[0054] The energy supply switching module outputs the ultimately determined optimal energy supply mode instruction to the system feedback module for triggering a corresponding energy supply control operation.

[0055] As a preferred solution of the intelligent power consumption management system for IoT devices of the present invention, wherein: the reinforcement learning method is used to evaluate the execution effect of the optimal energy supply mode instruction to obtain system parameter optimization suggestions, and the specific steps are as follows:

[0056] Build a reinforcement learning model, including the state space, action space, and reward function;

[0057] The state space is represented as a triplet, which includes the energy supply mode, the energy storage state and the environmental energy input;

[0058] The action space is represented as an executable parameter adjustment scheme;

[0059] Update the strategy through the Q-learning algorithm;

[0060] Output parameter optimization suggestions, wherein the parameter optimization suggestions include a dynamic sleep threshold adjustment amount, an energy consumption weight coefficient adjustment amount, and a power supply switching threshold adjustment amount.

[0061] In a second aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the intelligent power consumption management system for Internet of Things devices as described in the first aspect of the present invention is implemented.

[0062] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, any step of the intelligent power consumption management system for Internet of Things devices as described in the first aspect of the present invention is implemented.

[0063] ​The beneficial effects of the present invention are as follows: the dynamic sleep threshold is calculated based on activity feature data, the optimized threshold parameters are output through boundary condition judgment, and the forced setting of boundary conditions solves the failure problem of traditional adaptive algorithms in extreme cases. The output dynamic threshold is directly related to the device activity, providing key input parameters for subsequent multi-objective optimization. The environmental parameters and power consumption scheme are processed through the LSTM neural network to predict the energy supply situation. The unique light + vibration dual energy prediction model greatly improves the energy prediction accuracy of field equipment, and the design of synchronizing the prediction cycle with the power consumption distribution forms a closed loop of prediction-distribution-verification, thereby maximizing the energy storage utilization rate in solar energy equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0065] Figure 1 This is a schematic diagram of the intelligent power consumption management system for IoT devices in Example 1. DETAILED DESCRIPTION

[0066] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0067] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0068] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0069] Example 1, with reference to Figure 1 , which is the first embodiment of the present invention, provides an intelligent power consumption management system for IoT devices, including:

[0070] Behavior collection module, dynamic adjustment module, energy consumption optimization module, energy prediction module, energy supply switching module and system feedback module;

[0071] Behavior collection module, used to monitor the operating status of the equipment and obtain equipment activity feature data;

[0072] Furthermore, a timed interrupt sampling mechanism is set up to collect equipment operation status data at fixed time intervals;

[0073] The equipment operation status data includes the sending / receiving status flag of the communication module, the real-time load rate of the central processing unit and the data acquisition flag of the sensor module;

[0074] Within the preset sliding time window, the activity characteristic value A is calculated based on the collected equipment operation status data t , the expression is:

[0075]

[0076] Among them, α and β are preset weighting coefficients, N active Indicates the number of active sampling points in the window, N total Indicates the total number of sampling points in the window, L i represents the normalized value of the CPU load rate at the i-th sampling point;

[0077] The calculated activity characteristic value A t Store in a ring buffer and calculate the mean and standard deviation of the historical activity feature values ​​in the buffer;

[0078] When the buffer is full, the device activity feature data is output, which includes the current activity feature value, the mean of historical activity features, and the standard deviation of historical activity features;

[0079] It should be noted that the behavior collection module realizes real-time monitoring of the device operation status through a timed interrupt sampling mechanism, and uses a circular buffer to store historical activity feature data. The design not only ensures the continuity and accuracy of data collection, but also effectively manages memory usage and avoids data overflow problems caused by long-term operation.

[0080] A dynamic adjustment module is used to process device activity feature data using an adaptive learning algorithm to obtain dynamic sleep threshold parameters;

[0081] Furthermore, the dynamic sleep threshold parameter is calculated based on the received device activity characteristic data, and the expression is:

[0082]

[0083] Among them, T base represents the basic dormancy threshold, k represents the adjustment coefficient, A t represents the current activity characteristic value, represents the mean value of historical activity characteristics, σ Arepresents the standard deviation of historical activity characteristics;

[0084] The calculated dynamic sleep threshold parameter T sleep Make boundary condition judgments and enforce them when specific conditions are met;

[0085] When A t >μ A +2σ A When T sleep Forced to 0.5 seconds;

[0086] When A t <μ A -2σ A When T sleep Forced to 10 seconds;

[0087] Outputting the finalized dynamic sleep threshold parameters to the energy consumption optimization module;

[0088] It should be noted that the adaptive learning algorithm in the dynamic adjustment module can automatically adjust the sleep threshold according to the actual operating conditions of the device, thereby reducing energy consumption as much as possible while ensuring the response speed of the device. In addition, by performing boundary condition judgment and forcible setting on the calculated dynamic sleep threshold parameters, it can prevent the device from entering an inappropriate sleep state in extreme cases, thereby ensuring the stability and reliability of the system.

[0089] Energy consumption optimization module, which is used to apply the dynamic sleep threshold parameters using a multi-objective optimization method to obtain the power consumption allocation plan for each functional module;

[0090] Furthermore, a multi-objective function is defined, and the optimal power consumption allocation scheme is calculated based on the constraints configured in the multi-objective optimization function. The expression is:

[0091] Minimizew1×P comm +w2×P comp +w3×P sense ;

[0092] P comm +P comp +P sense ≤P max ;

[0093] T response ≤T sleep ×2;

[0094] P comm ≥P comm_min ;

[0095] P comp ≥P comp_min ;

[0096] P sense ≥P sense_min ;

[0097] Among them, w1, w2, w3 are the weight coefficients of communication, computing and sensing modules respectively, P max is the upper limit of the total power consumption of the system, T response is the current response time of the system, P comm_min , P comp_min , P sense_min are the minimum operating power consumption of each module respectively;

[0098] The linear programming algorithm is used to solve the above optimization problem and obtain the power consumption distribution value of each functional module. The power consumption distribution value of each functional module includes the power consumption distribution value of the communication module P comm_opt , calculate the module allocation power consumption P comp_opt And the sensor module allocates power consumption P sense_opt ;

[0099] The upper limit of the total system power consumption needs to be dynamically adjusted according to the current remaining power of the device. The expression is:

[0100]

[0101] Among them, P base Basic power consumption limit, E remain Current remaining power, E threshold The power threshold, when it is lower than this value, the power consumption is limited. full The device's full power level and the k adjustment factor, used to control the sensitivity of power consumption limits.

[0102] It should be noted that the energy consumption optimization module uses a multi-objective optimization method to comprehensively consider the power consumption allocation problem of the three main functional modules of communication, computing and sensing, aiming to find a working mode that can meet the system performance requirements and maximize energy savings. At the same time, the module dynamically adjusts the upper limit of the system's total power consumption according to the current remaining power, so that the entire system can flexibly respond to various application scenarios under different power conditions, thereby improving the system's adaptability and robustness.

[0103] Energy prediction module, which is used to analyze the power consumption allocation scheme based on the neural network model and the environmental parameters to obtain the energy supply prediction result;

[0104] Furthermore, the energy prediction module collects current environmental parameters, including light intensity L, ambient temperature T, and vibration intensity V;

[0105] Input the received power consumption allocation plan and environmental parameters into the pre-trained LSTM neural network model;

[0106] The structure of the LSTM neural network model includes input layer, hidden layer and output layer;

[0107] Collect historical data samples, including historical power consumption allocation plans, environmental parameters within the corresponding time period, and actual measured energy input data;

[0108] Normalize the collected data;

[0109] The Adam optimizer is used to train the model. The loss function is defined as the sum of the mean square error (MSE) and the L2 regularization term. The expression is:

[0110] Loss = MSE + λ × L2 regularization;

[0111] Among them, λ is the regularization coefficient, which is used to control the balance between model complexity and generalization ability;

[0112] The input layer contains 5 neurons, corresponding to P comm_opt 、P comp_opt 、P sense_opt , light intensity and ambient temperature;

[0113] The hidden layer consists of two LSTM layers, each containing 32 neurons;

[0114] The output layer contains 3 neurons, which output the following prediction results:

[0115] Solar energy input forecast value E in the next 1 hour solar ;

[0116] Vibration energy input prediction value E in the next 1 hour vib ;

[0117] Predicted value of total environmental energy input E total ;

[0118] The energy supply prediction result E is calculated by the LSTM neural network model total , the expression is:

[0119] E total =E solar +E vib ;

[0120] in:

[0121] E solar =f solar (P comm_opt ,P comp_opt ,P sense_opt ,L,T);

[0122] E vib =f vib(P comm_opt ,P comp_opt ,P sense_opt ,V);

[0123] The final calculated total environmental energy input prediction result E total Output to the energy supply switching module for subsequent energy scheduling and power supply mode decision-making;

[0124] The energy supply forecasting cycle of the energy prediction module is consistent with the update cycle of the power consumption allocation plan generated by the energy consumption optimization module, which is used to ensure that the system can achieve real-time matching between energy prediction and power consumption allocation during dynamic adjustment;

[0125] It should be noted that the energy prediction module predicts environmental energy input based on the LSTM neural network model. This method is suitable for processing data with time series characteristics, such as light intensity, temperature changes, etc. By pre-training the model and combining it with real-time analysis of actual environmental parameters, it can provide accurate energy supply prediction results and help the system make more reasonable energy supply decisions. In addition, the design of the synchronized update cycle of this module and the energy consumption optimization module further enhances the overall coordination and response efficiency of the system.

[0126] The energy supply switching module is used to respond to the energy supply prediction results using a priority strategy to obtain the optimal energy supply mode instruction;

[0127] Furthermore, the energy supply mode is determined based on the priority strategy;

[0128] The priority strategies are as follows:

[0129] When E total ≥E threshold_hiht When the system is in the state of emergency, the environment energy supply system is used first, and the excess energy is stored in the energy storage unit. At the same time, the energy supply mode instruction is output as the environment energy priority mode;

[0130] When E threshold_low ≤E total <E threshold_high When the power supply mode is combined with ambient energy and energy storage, the output power supply mode instruction is a hybrid power supply mode, in which the ambient energy usage ratio is calculated according to the formula:

[0131]

[0132] When E total <E threshold_low And S≥S critical When the power supply is in the energy storage mode, the energy storage priority mode is used first.

[0133] When E total <E threshold_low And S <Scritical When the low power consumption operation mechanism is started, the output power supply mode instruction is the emergency energy saving mode;

[0134] The energy supply switching module outputs the final determined optimal energy supply mode instruction to the system feedback module to trigger the corresponding energy supply control operation;

[0135] It should be noted that the energy supply switching module adopts a priority strategy to determine the optimal energy supply mode instructions. The strategy fully considers the different characteristics of energy demand in different scenarios to ensure that the normal operation of the system can be maintained under any circumstances. Especially in emergency energy-saving mode, the system will automatically switch to a low-power operation state to extend the working time of the equipment until sufficient external energy supply is obtained or the energy storage is restored to a safe level.

[0136] The system feedback module is used to evaluate the execution effect of the optimal energy supply mode instruction using reinforcement learning methods and obtain parameter optimization suggestions;

[0137] Going further, build a reinforcement learning model, which includes state space, action space and reward function;

[0138] The state space is represented as a triplet, which includes the energy supply mode, energy storage state and environmental energy input;

[0139] The action space is represented as executable parameter adjustment schemes;

[0140] The reward function R is defined as follows:

[0141]

[0142] Among them, E actual is the actual energy consumption data, T response is the system response time, R completion is the task completion rate w1, w2, w3 are weight coefficients, E max is the maximum expected energy input;

[0143] The strategy is updated through the Q-learning algorithm, and the expression is:

[0144]

[0145] Among them, α is the learning rate, which controls the update step size, γ is the discount factor, which is used to measure the importance of future rewards, s is the current state, a is the action taken in the current state, r is the immediate reward, s ′ Next state, a ′ Optional actions in the next state;

[0146] Output parameter optimization suggestions, including dynamic sleep threshold adjustment, energy consumption weight coefficient adjustment, and power supply switching threshold adjustment;

[0147] It should be noted that the system feedback module uses reinforcement learning methods to evaluate the execution effect of the optimal energy supply mode instructions, and based on this, it puts forward parameter optimization suggestions, realizing the transformation from passive response to active optimization. This method not only helps to improve the system's immediate performance, but also continuously improves its own decision-making ability through a continuous learning process, and ultimately achieves the goal of improving overall energy efficiency. The system can maintain an efficient and stable operating state in a complex and changing application environment.

[0148] This embodiment also provides a computer device suitable for the intelligent power consumption management system of Internet of Things devices, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the intelligent power consumption management system of Internet of Things devices proposed in the above embodiment.

[0149] The computer device may be a terminal, comprising a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner may be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. The display screen of the computer device may be a liquid crystal display or an electronic ink display screen, and the input device of the computer device may be a touch layer covering the display screen, or a button, trackball or touchpad provided on the housing of the computer device, or an external keyboard, touchpad or mouse.

[0150] This embodiment also provides a storage medium having a computer program stored thereon, which, when executed by a processor, implements the intelligent power consumption management system of the Internet of Things device as proposed in the above embodiment; the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0151] In summary, the present invention calculates the dynamic sleep threshold based on activity feature data, and outputs the optimized threshold parameters through boundary condition judgment. The forced setting of boundary conditions solves the failure problem of traditional adaptive algorithms in extreme cases. The output dynamic threshold is directly related to the device activity, which provides key input parameters for subsequent multi-objective optimization. The environmental parameters and power consumption scheme are processed through the LSTM neural network to predict the energy supply situation. The unique light + vibration dual energy prediction model greatly improves the energy prediction accuracy of field equipment, and the design of synchronizing the prediction cycle with the power consumption distribution forms a closed loop of prediction-distribution-verification, thereby maximizing the energy storage utilization rate in solar energy equipment.

[0152] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. An intelligent power consumption management system for IoT devices, characterized by: include: Behavior collection module, dynamic adjustment module, energy consumption optimization module, energy prediction module, energy supply switching module and system feedback module; The behavior collection module is used to monitor the operating status of the device and obtain device activity feature data; The dynamic adjustment module is used to process the device activity feature data using an adaptive learning algorithm to obtain a dynamic sleep threshold parameter; The energy consumption optimization module is used to apply a multi-objective optimization method to the dynamic sleep threshold parameters to obtain a power consumption allocation plan for each functional module; The energy prediction module is used to analyze the power consumption allocation scheme in combination with environmental parameters based on a neural network model to obtain an energy supply prediction result; The energy supply switching module is used to respond to the energy supply prediction result using a priority strategy to obtain an optimal energy supply mode instruction; The system feedback module is used to evaluate the execution effect of the optimal energy supply mode instruction using a reinforcement learning method to obtain parameter optimization suggestions.

2. The intelligent power consumption management system for IoT devices according to claim 1, wherein: The specific steps of monitoring the equipment operation status and obtaining the equipment activity characteristic data are as follows: Set up a timed interrupt sampling mechanism to collect equipment operating status data at fixed time intervals; The equipment operation status data includes the sending / receiving status flag of the communication module, the real-time load rate of the central processing unit and the data acquisition flag of the sensor module; Calculate activity feature values ​​based on the collected equipment operating status data within a preset sliding time window; The calculated activity characteristic values ​​are stored in a ring buffer, and the mean and standard deviation of the historical activity characteristic values ​​in the buffer are calculated; When the buffer is full, device activity feature data is output, where the device activity feature data includes a current activity feature value, a historical activity feature mean, and a historical activity feature standard deviation.

3. The intelligent power consumption management system for IoT devices according to claim 2, wherein: The adaptive learning algorithm is used to process the device activity feature data to obtain the dynamic sleep threshold parameter. The specific steps are as follows: Calculating a dynamic sleep threshold parameter based on the received device activity characteristic data; The calculated dynamic sleep threshold parameter T sleep Make boundary condition judgments and enforce them when specific conditions are met; When A t >μ A +2σ A When T sleep Forced to 0.5 seconds; When A t <μ A -2σ A When T sleep Forced to 10 seconds; The finalized dynamic sleep threshold parameter is output to the energy consumption optimization module.

4. The intelligent power consumption management system for IoT devices according to claim 3, wherein: The multi-objective optimization method is applied to the dynamic sleep threshold parameters to obtain the power consumption allocation scheme of each functional module. The specific steps are as follows: Define a multi-objective function and calculate the optimal power consumption allocation solution based on the constraints configured in the multi-objective optimization function; The linear programming algorithm is used to solve the above optimization problem to obtain the power consumption distribution value of each functional module, which includes the power consumption distribution value of the communication module P comm_opt , calculate the module allocation power consumption P comp_opt And the sensor module allocates power consumption P sense_opt .

5. The intelligent power consumption management system for IoT devices according to claim 4, wherein: The power consumption allocation scheme is analyzed based on the neural network model in combination with environmental parameters to obtain energy supply prediction results. The specific steps are as follows: The energy prediction module collects current environmental parameters, which include light intensity L, ambient temperature T, and vibration intensity V; Input the received power consumption allocation plan and environmental parameters into the pre-trained LSTM neural network model; The structure of the LSTM neural network model includes an input layer, a hidden layer and an output layer; The input layer contains 5 neurons, corresponding to P comm_opt 、P comp_opt 、P sense_opt , light intensity and ambient temperature; The hidden layer consists of two LSTM layers, each containing 32 neurons; The output layer contains 3 neurons, which output the following prediction results respectively: Solar energy input forecast value E in the next 1 hour solar ; Vibration energy input prediction value E in the next 1 hour vib ; Predicted value of total environmental energy input E total ; The energy supply prediction result E is calculated by the LSTM neural network model total ; The final calculated total environmental energy input prediction result E total The output is sent to the energy supply switching module for subsequent energy scheduling and power supply mode decision-making.

6. The intelligent power consumption management system for IoT devices according to claim 5, characterized in that: The LSTM neural network model is modeled and optimized through the following training process: Collecting historical data samples, wherein the historical data samples include historical power consumption allocation plans, environmental parameters within a corresponding time period, and actually measured energy input data; Normalize the collected data; The model is trained using the Adam optimizer, and the loss function is defined as the sum of the mean square error (MSE) and the L2 regularization term. The energy supply prediction cycle of the energy prediction module is consistent with the update cycle of the power consumption allocation scheme generated by the energy consumption optimization module, so as to ensure that the system achieves real-time matching between energy prediction and power consumption allocation during dynamic adjustment.

7. The intelligent power consumption management system for IoT devices according to claim 6, wherein: The priority strategy is used to respond to the energy supply prediction result to obtain the optimal energy supply mode instruction, and the specific steps are: Determine the energy supply mode based on the priority strategy; The priority strategy is specifically as follows: When E total ≥E threshold_high When the system is in the state of emergency, the environment energy supply system is used first, and the excess energy is stored in the energy storage unit. At the same time, the energy supply mode instruction is output as the environment energy priority mode; When E threshold_low ≤E total <R threshold_high When the power supply mode is combined with ambient energy and energy storage, the output power supply mode instruction is a hybrid power supply mode; When E total <E threshold_low And S≥S critical When the power supply is in the energy storage mode, the energy storage priority mode is used first. When E total <E threshold_low And S critical When the low power consumption operation mechanism is started, the output power supply mode instruction is the emergency energy saving mode;​ The energy supply switching module outputs the ultimately determined optimal energy supply mode instruction to the system feedback module for triggering a corresponding energy supply control operation.

8. The intelligent power consumption management system for IoT devices according to claim 7, wherein: The reinforcement learning method is used to evaluate the execution effect of the optimal energy supply mode instruction to obtain system parameter optimization suggestions. The specific steps are as follows: Build a reinforcement learning model, including the state space, action space, and reward function; The state space is represented as a triplet, which includes the energy supply mode, the energy storage state and the environmental energy input; The action space is represented as an executable parameter adjustment scheme; Update the strategy through the Q-learning algorithm; Output parameter optimization suggestions, wherein the parameter optimization suggestions include a dynamic sleep threshold adjustment amount, an energy consumption weight coefficient adjustment amount, and a power supply switching threshold adjustment amount.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the intelligent power consumption management system for Internet of Things devices according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the intelligent power consumption management system for Internet of Things devices according to any one of claims 1 to 8 are implemented.

Images