Intelligent equipment control system for classroom

Abstract

The invention relates to the technical field of equipment intelligent control, and particularly discloses an equipment intelligent control system for a classroom, and the system comprises a state judgment module which comprises a multi-mode sensing device, a data fusion unit and a state judgment algorithm; the distributed execution module comprises a plurality of execution units and a communication network; and the central management unit performs data interaction with the state judgment module, the equipment control strategy generation module and the distributed execution module through the wireless communication module, and performs data synchronization with an educational administration system and a campus security monitoring system. Through intelligent coordination of multiple modules, the automation level of classroom equipment management is improved, powerful guarantee is provided for energy conservation and safety, and the teaching environment is optimized.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent device control, and particularly to an intelligent control system for classroom devices. Background Art

[0002] With the rapid development of intelligent technologies, the education field has gradually introduced intelligent devices and control systems to improve teaching efficiency, optimize resource allocation, and ensure the safety and comfort of the teaching environment. The traditional management method of classroom devices relies on manual operation, and cannot automatically adjust the device status according to environmental changes in real time, resulting in energy waste, low environmental comfort, and great management difficulty. Therefore, developing an intelligent device control system has become the key to improving classroom environment management.

[0003] Existing classroom device control systems mostly focus on a single environmental factor or device control strategy, such as automatic adjustment of temperature, lighting, or air conditioning, etc., but often lack comprehensive consideration of multiple environmental factors (such as temperature, light, seat occupancy, etc.). At the same time, existing systems also have deficiencies in the collaborative work and energy management among devices, and it is difficult to achieve intelligent scheduling and optimal control.

[0004] Therefore, designing an intelligent control system for classroom devices has become an urgent problem to be solved currently. Summary of the Invention

[0005] In order to overcome the above technical problems existing in the prior art, the embodiments of the present invention provide an intelligent control system for classroom devices. Through the intelligent coordination of multiple modules, it not only improves the automation level of classroom device management, but also provides a strong guarantee for energy conservation and safety, and optimizes the teaching environment.

[0006] To achieve the above object, an embodiment of the present invention provides an intelligent control system for classroom equipment. The system includes: a state determination module, which includes a multi-modal sensing device, a data fusion unit, and a state determination algorithm; the data fusion unit fuses the data collected by the multi-modal sensing device to generate a classroom environment information sequence; the state determination algorithm determines the classroom usage status based on the classroom environment information sequence, in combination with the course scheduling data of the academic affairs system and the real-time data of the campus security monitoring system; a device control strategy generation module based on deep learning, which includes a deep learning model and a control strategy optimization algorithm; the deep learning model is trained based on historical data, with the input being the classroom environment information sequence and the classroom usage status, and the output being the control strategy of the device; the control strategy optimization algorithm optimizes the control strategy of the device based on the output of the deep learning model, in combination with the energy consumption model and the security model, to generate an optimized control strategy; a distributed execution module, which includes multiple execution units and a communication network; the execution unit: each execution unit corresponds to a device in the classroom. After receiving the control command, the execution unit interacts through a wireless communication module to achieve the switching and adjustment of the device; the wireless communication module: adopts a half-duplex communication protocol based on power line to support the cooperative control between devices; the central management unit conducts data interaction with the state determination module, the device control strategy generation module, and the distributed execution module through the wireless communication module. The central management unit also synchronizes data with the academic affairs system and the campus security monitoring system.

[0007] Preferably, the multi-modal sensing device includes an infrared sensor, an ultrasonic sensor, a light intensity sensor, and a weight sensor distributed in the classroom; among them, the infrared sensor is installed at the four corners and the door of the classroom to detect the heat source distribution in the classroom; the ultrasonic sensor is installed on the surrounding walls of the classroom to detect the movement of objects in the classroom; the light intensity sensor is installed on the ceiling of the classroom to detect the light intensity in the classroom; the weight sensor is installed under each seat in the classroom to detect the occupancy of the seat.

[0008] Preferably, the data fusion unit fuses the data collected by the multi-modal sensing device to generate a classroom environment information sequence (T, U, L, G).

[0009] Data fusion formula:

[0010] S = αT + βU + γL + δG;

[0011] Wherein, S represents the comprehensive classroom environment information, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, and α, β, γ, and δ are the weight coefficients of each information respectively.

[0012] Preferably, the status determination algorithm determines the classroom usage status based on the classroom environment information sequence (T, U, L, G), combines the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system, and passes through a preset decision tree model.

[0013] Decision tree model: Generate a decision tree model based on the heat source information, object movement information, light intensity information, and seat occupancy information of the classroom, as well as the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system;

[0014] The training data set of the decision tree model includes the historical classroom environment information sequence and the corresponding classroom usage status. By training, the parameters of the model are optimized so that the model can accurately determine the classroom usage status.

[0015] Preferably, the deep learning model is trained based on historical data. The input is the classroom environment information sequence (T, U, L, G) and the classroom usage status, and the output is the control strategy of the device;

[0016] Deep learning model training formula:

[0017] f(S, O) → C;

[0018] Among them, f is the deep learning model, S is the comprehensive classroom environment information, O is the classroom usage status, and C is the control strategy of the device.

[0019] Preferably, the control strategy optimization algorithm is based on the output of the deep learning model, combines the energy consumption model and the security model, optimizes the control strategy of the device, and generates an optimized control strategy;

[0020] Energy consumption model: Generate an energy consumption data set based on the historical classroom environment information sequence, the actual control strategy of the device, as well as the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system;

[0021] Construct an energy consumption model through the energy consumption data set to predict the energy consumption situation of the device;

[0022] Security model: Generate a security data set based on the historical classroom environment information sequence, the actual control strategy of the device, as well as the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system;

[0023] Construct a security model through the security data set to evaluate the security of the device control strategy;

[0024] Control strategy optimization algorithm formula:

[0025] C final = argmin(λ1·E + λ2·S);

[0026] Among them, C finalLet \(C\) represent the optimized control strategy, \(E\) represent the energy consumption, \(S\) represent the security, and \(\lambda_1\) and \(\lambda_2\) be the weight coefficients of energy consumption and security respectively.

[0027] Preferably, the energy consumption model adopts a linear regression algorithm, and the expression of the linear regression algorithm is:

[0028] \(E = \beta_0+\beta_1T+\beta_2U+\beta_3L+\beta_4G+\beta_5C+\beta_6S\) c +\(\beta_7S\) s +\(\epsilon\);

[0029] Among them, \(E\) represents the energy consumption of the device, \(T\) represents the heat source information, \(U\) represents the object movement information, \(L\) represents the light intensity information, \(G\) represents the seat occupancy information, \(C\) is the control strategy of the device, \(S\) c represents the course scheduling data of the educational administration system, \(S\) s represents the real-time data of the campus security monitoring system, \(\beta_0\) represents the intercept term of the model, \(\beta_1\) to \(\beta_7\) are the weights of the model, and \(\epsilon\) represents the error term, that is, the difference between the actual energy consumption and the model-predicted energy consumption. The weights of the model are optimized by the least squares method.

[0030] Preferably, the security model adopts a logistic regression model, and the expression of the logistic regression model is:

[0031]

[0032] Among them, \(P\) represents the security probability of the device control strategy, \(T\) represents the heat source information, \(U\) represents the object movement information, \(L\) represents the light intensity information, \(G\) represents the seat occupancy information, \(C\) is the control strategy of the device, \(S\) c represents the course scheduling data of the educational administration system, \(S\) s represents the real-time data of the campus security monitoring system, and \(\alpha_0\) to \(\alpha_7\) are the weights of the model. The weights of the model are optimized by the gradient descent method.

[0033] Preferably, the execution unit receives a control command from the device control strategy generation module, and the control command includes the on / off state and adjustment parameters of the device;

[0034] Execution unit control command generation formula:

[0035] \(C\) unit \(= f(C\) final ) where \(C\) unit represents the specific control signal generated by the execution unit, \(C\) final represents the optimized control strategy, and \(f\) represents any function that converts \(C\) final to \(C\) unit .

[0036] Preferably, the central management unit synchronizes data with the educational administration system and the campus security monitoring system through the wireless communication module to obtain the latest course scheduling data and monitoring data;

[0037] Central management unit data synchronization formula:

[0038] D sync =f(S c ,S s );

[0039] Among them, D sync represents the synchronized data, S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system.

[0040] Through the technical solution provided by the present invention, the present invention has at least the following technical effects:

[0041] 1. Through this intelligent control system, the automatic adjustment of the equipment in the classroom can be realized, thereby improving the energy utilization efficiency, optimizing the learning environment, and enhancing the safety. At the same time, the system can flexibly adjust the control strategy according to the real-time data and historical data to adapt to different usage scenarios and requirements, ensuring the efficient operation of the equipment, the comfortable environment, and the safety and reliability.

[0042] 2. Through the collaborative work of multi-modal sensors, the system can real-time sense environmental factors such as personnel activities, lighting conditions, and temperature distribution in the classroom, and intelligently adjust the equipment based on these data. For example, the system can automatically adjust the air conditioner and lights according to the seat occupancy situation, improving the comfort level while saving energy.

[0043] 3. Through the weighted fusion processing of environmental information, the system can dynamically adjust the equipment status according to the actual needs of the classroom. When the lighting in the classroom is sufficient and there is no personnel activity, devices such as lights and air conditioners can be automatically turned off or the power can be reduced to achieve energy-saving effects. This not only reduces energy waste but also extends the service life of the equipment.

[0044] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent specific implementation part. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. They are used together with the following specific implementation to explain the embodiments of the present invention, but do not constitute a limitation to the embodiments of the present invention. In the drawings:

[0046] Figure 1 is the general block diagram of an equipment intelligent control system for a classroom provided by an embodiment of the present application;

[0047] Figure 2 Block diagram of a state determination module of an equipment intelligent control system for a classroom provided by an embodiment of the present application;

[0048] Figure 3 Block diagram of a multimodal perception device of an equipment intelligent control system for a classroom provided by an embodiment of the present application. Detailed implementation manners

[0049] The following details the specific implementation manners of the embodiments of the present invention with reference to the accompanying drawings. It should be understood that the specific implementation manners described herein are only for the purpose of illustration and explanation of the embodiments of the present invention, and are not used to limit the embodiments of the present invention.

[0050] The terms "system" and "network" in the embodiments of the present invention can be used interchangeably. "A plurality of" means two or more. In view of this, in the embodiments of the present invention, "a plurality of" can also be understood as "at least two". "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " generally represents an "or" relationship between the associated objects before and after, unless otherwise specified. In addition, it should be understood that in the description of the embodiments of the present invention, words such as "first" and "second" are only used for the purpose of distinguishing descriptions, and cannot be understood as indicating or implying relative importance, nor can they be understood as indicating or implying order.

[0051] Please refer to Figures 1 - 3, an embodiment of the present invention provides an intelligent control system for classroom devices. The system includes: a status determination module: including a multimodal perception device, a data fusion unit, and a status determination algorithm; the data fusion unit fuses the data collected by the multimodal perception device to generate a classroom environment information sequence; the status determination algorithm: based on the classroom environment information sequence, combined with the course scheduling data of the academic affairs system and the real-time data of the campus security monitoring system, determines the classroom usage status; a device control strategy generation module based on deep learning: including a deep learning model and a control strategy optimization algorithm; the deep learning model: trained based on historical data, with the input being the classroom environment information sequence and the classroom usage status, and the output being the control strategy of the device; the control strategy optimization algorithm: based on the output of the deep learning model, combined with the energy consumption model and the security model, optimizes the control strategy of the device to generate an optimized control strategy; a distributed execution module: including multiple execution units and a communication network; the execution unit: each execution unit corresponds to a device in the classroom. After receiving the control command, the execution unit interacts through the wireless communication module to achieve the on / off and adjustment of the device; the wireless communication module: adopts a half-duplex communication protocol based on power lines to support the collaborative control between devices; the central management unit: conducts data interaction with the status determination module, the device control strategy generation module, and the distributed execution module through the wireless communication module. The central management unit also synchronizes data with the academic affairs system and the campus security monitoring system.

[0052] In a possible implementation manner, multimodal perception devices are arranged in the classroom, including infrared sensors, ultrasonic sensors, light intensity sensors, weight sensors, etc., to collect environmental data in real time. The output signals of these sensors are processed by the data fusion unit to generate a comprehensive classroom environment information sequence. This information includes temperature, object movement, light intensity, and seat occupancy, etc. The data fusion unit fuses the data from different sensors to form a complete environment information sequence, providing a basis for subsequent status determination.

[0053] Furthermore, based on the classroom environment information sequence obtained from the data fusion unit, the status determination algorithm will combine the course scheduling data provided by the academic affairs system and the real-time data of the campus security monitoring system to judge the classroom usage status. The course scheduling data can inform the system whether there is a course in progress currently, as well as the start and end times of the course; the security monitoring system provides the security status data inside and outside the classroom. By integrating this information, the status determination algorithm can determine whether the classroom is in use, whether the device needs to be started, and the current usage mode (such as in class, after school, etc.).

[0054] Furthermore, based on historical data (such as past classroom environment data and equipment control data), a deep learning model is trained to obtain an equipment control strategy. The input is a sequence of classroom environment information and the classroom usage status, and the model outputs the corresponding equipment control strategy. For example, according to the temperature, light intensity, seat occupancy, etc. in the classroom, the model may generate control instructions such as corresponding air conditioner temperature settings and light brightness adjustments. This process utilizes the training of historical data and can automatically generate adaptive control strategies according to different environmental conditions.

[0055] Furthermore, the preliminary control strategy output by the deep learning model will enter the control strategy optimization algorithm module. This module combines an energy consumption model and a safety model to optimize the control strategy. The energy consumption model is used to evaluate the energy consumption of equipment under different control strategies, and the optimization algorithm tries to reduce energy consumption while ensuring the normal operation of the equipment. The safety model will consider the safety of equipment operations to ensure that any control command will not cause equipment failures or unsafe situations. Finally, control commands will be generated through the optimized control strategy.

[0056] Furthermore, the optimized control commands are transmitted to each execution unit through a wireless communication module. Each execution unit corresponds to a specific device in the classroom, such as an air conditioner, lights, a projector, etc. The execution units perform cooperative control with other execution units through a power line-based half-duplex communication protocol to achieve the switching and adjustment of devices. The design of the wireless communication module enables each device to perform intelligent cooperation through the same network, ensuring that the devices in the classroom can automatically adjust according to changes in the classroom environment.

[0057] Furthermore, the central management unit is the control center of the entire system, and it conducts data interaction with the status determination module, the equipment control strategy generation module, and the distributed execution module through the wireless communication module to coordinate the work of each module. The central management unit also synchronizes data with the teaching affairs system and the campus security monitoring system to ensure that the system can timely grasp information such as class scheduling and security, so as to perform timely equipment regulation. For example, when the class schedule in the classroom changes, the central management unit will adjust the equipment operation strategy to ensure that the equipment operates optimally under different usage modes.

[0058] Through the intelligent coordination of multiple modules, this equipment intelligent control system not only improves the automation level of classroom equipment management, but also provides strong guarantees for energy conservation and safety, and optimizes the teaching environment.

[0059] In an embodiment of the present invention, the multimodal perception device includes infrared sensors, ultrasonic sensors, light intensity sensors, and weight sensors distributed in the classroom. Among them, the infrared sensors are installed at the four corners and the door of the classroom and are used to detect the heat source distribution in the classroom. The ultrasonic sensors are installed on the surrounding walls of the classroom and are used to detect the movement of objects in the classroom. The light intensity sensors are installed on the ceiling of the classroom and are used to detect the light intensity in the classroom. The weight sensors are installed under each seat in the classroom and are used to detect the occupancy of the seats.

[0060] In a possible implementation manner, the signals of all sensors are finally converged to the data fusion unit of the system. The data fusion unit comprehensively processes the signals from the infrared, ultrasonic, light intensity, and weight sensors, fuses the data from different sensors, and generates a complete classroom environment information sequence. The fused data is used to support the state determination of the system and the generation of device control strategies, and finally realizes the intelligent adjustment of the device.

[0061] Specifically, the infrared sensors are installed at the four corners and the door of the classroom and are mainly used to detect the heat source distribution in the classroom. By monitoring the heat change, the infrared sensors can detect the position and activity status of the people in the classroom. When someone enters or moves in the classroom, the infrared sensors will sense the change of the heat source and transmit the signal to the data fusion unit. The data fusion unit judges the personnel dynamics in the classroom based on these signals and provides input data for the device control strategy.

[0062] Furthermore, the ultrasonic sensors are installed on the surrounding walls of the classroom and are mainly used to detect the movement of objects in the classroom. The ultrasonic sensors emit ultrasonic waves and receive the reflected waves to judge the presence and position of the objects. When an object (such as a person, furniture, etc.) moves in the classroom, the ultrasonic sensors will detect the change of the reflected waves, thereby obtaining the dynamic information of the objects in the classroom. The sensors transmit this information to the data fusion unit to further assist in judging whether there are people in the classroom, and then adjust the working state of the devices (such as air conditioners, lights, etc.).

[0063] Furthermore, the light intensity sensors are installed on the ceiling of the classroom and are mainly used to detect the light intensity in the classroom. By monitoring the light change in the classroom in real time, the sensors can evaluate the natural light conditions in the classroom. When the light in the classroom is insufficient, the system can automatically adjust the artificial light source (such as the brightness of the lights) to ensure that the light in the classroom always remains within a comfortable range. The signals of the light intensity sensors are also transmitted to the data fusion unit as an important basis for optimizing the device control strategy.

[0064] Further, weight sensors are installed under each seat in the classroom to detect the occupancy of the seats. When a student sits on a seat, the weight sensor will sense the change in the weight of the seat and transmit the signal to the data fusion unit. The system determines whether there are students attending classes in the classroom based on the occupancy of each seat, and then decides whether to start or adjust devices such as air conditioners and lights in the classroom.

[0065] In the embodiment of the present invention, the data fusion unit fuses the data collected by the multi-modal perception device to generate a classroom environment information sequence (T, U, L, G);

[0066] Data fusion formula:

[0067] S = αT + βU + γL + δG;

[0068] Wherein, S represents the comprehensive classroom environment information, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, and α, β, γ, and δ are the weight coefficients of each information respectively.

[0069] In a possible implementation manner, the multi-modal perception device collects various types of information in the classroom environment through different sensors, and these information are respectively:

[0070] T (heat source information): The temperature sensor installed in the classroom measures the temperature data in the classroom in real time to judge the comfort level in the classroom and the cold and hot state of the environment.

[0071] U (object movement information): Collected by ultrasonic sensors, it can detect the movement information of objects (such as people) in the classroom, and then reflect the activity of people in the classroom.

[0072] L (light intensity information): Collected by light intensity sensors, it detects the light intensity in the classroom in real time to judge whether it is necessary to adjust artificial lighting equipment to ensure suitable lighting conditions.

[0073] G (seat occupancy information): The weight sensor monitors the occupancy of the seats, reflecting whether each seat is occupied, so as to infer whether there are students in the classroom.

[0074] The data of these sensors are transmitted to the data fusion unit by wired or wireless means. The data fusion unit fuses various types of information (T, U, L, G) collected, and calculates the classroom environment information sequence S by weighted calculation according to the set weight coefficients (α, β, γ, δ). This information is used as the basis for subsequent device control.

[0075] Furthermore, each sensor (temperature sensor, ultrasonic sensor, light intensity sensor, weight sensor) continuously monitors the classroom environment and generates a real-time data stream. The collected data is transmitted to the data fusion unit via wired or wireless communication. The data fusion unit is the core of the system and is responsible for processing and integrating this information. In the data fusion unit, first, T, U, L, and G are processed and weighted and fused according to the set weight coefficients (α, β, γ, δ). The specific formula is: S = αT + βU + γL + δG;

[0076] Among them, α, β, γ, and δ represent the influence degree of each type of information on the comprehensive environment. This weight coefficient can be adjusted according to the requirements of system design. For example, temperature and seat occupancy information may be more important than light intensity and object movement information, so higher weights can be assigned to temperature and seat occupancy. The data fusion unit calculates the classroom environment information sequence S through this weighted method. This information synthesizes factors such as temperature, light, object movement, and seat occupancy in the classroom, providing an accurate basis for subsequent device control.

[0077] Furthermore, based on the classroom environment information sequence S, the system decides how to adjust the devices (such as air conditioners, lighting, etc.) in the classroom through an intelligent control algorithm. For example, when the S value indicates that the temperature in the classroom is too high and the light is insufficient, the system may adjust the air conditioner to lower the temperature and increase the light brightness.

[0078] Data fusion and weighted processing provide more accurate environmental monitoring and intelligent decision-making support for the device intelligent control system, not only improving the classroom management efficiency but also creating a more comfortable and energy-saving teaching environment.

[0079] In the embodiment of the present invention, the state determination algorithm is based on the classroom environment information sequence (T, U, L, G), combines the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system, and determines the classroom usage state through a preset decision tree model. Decision tree model: According to the heat source information, object movement information, light intensity information, and seat occupancy information of the classroom, as well as the course scheduling data of the educational administration system and the real-time data of the campus security monitoring system, a decision tree model is generated; the training data set of the decision tree model includes the historical classroom environment information sequence and the corresponding classroom usage state, and the parameters of the model are optimized through training so that the model can accurately determine the classroom usage state.

[0080] In a possible implementation manner, the system first collects the classroom environment information sequence (T, U, L, G) in real time through a multi-modal perception device, and these data are transmitted to the data fusion unit through sensors to form the real-time environment state of a classroom. At the same time, the educational administration system provides course scheduling data, and the campus security monitoring system provides real-time monitoring data. These information jointly provide input for the state determination algorithm.

[0081] T(Thermal source information): Reflects the temperature status of the classroom and affects comfort.

[0082] U(Object movement information): Detects the activities of people or objects in the classroom.

[0083] L(Light intensity information): Reflects the lighting environment in the classroom.

[0084] G(Seat occupancy information): Reflects the seat occupancy status in the classroom and helps to determine whether there are students present.

[0085] Meanwhile, the educational administration system provides the class schedule of the classroom (such as class time, course name, instructor, etc.), and the campus security monitoring system provides security data on whether there are abnormalities in the classroom (such as intrusion detection). This information is transmitted to the input end of the status determination algorithm through the data fusion unit and the interface.

[0086] Furthermore, the status determination algorithm determines the classroom usage status based on the decision tree model.

[0087] Specifically, in one embodiment, the decision tree model is trained through the historical classroom environment information sequence (T, U, L, G) and the corresponding usage status (such as "having class", "studying by oneself", "idle"). The decision tree recursively partitions the data according to the environmental information (such as temperature, light, etc.), and each branch represents a specific condition or decision point, and finally generates a tree-like structure for determining the classroom usage status.

[0088] Furthermore, the training data set of the decision tree model includes historical data (the classroom environment information sequence and the actual usage status). Through training, the decision tree model continuously optimizes the splitting rules and the selection of nodes, so that the model can make more accurate predictions based on the input environmental data and class schedule information in actual applications.

[0089] Furthermore, the decision tree model generates decision results based on the real-time input data, such as temperature, light intensity, seat occupancy, etc., and outputs the classroom usage status (for example, whether it is class time, whether there is someone in the classroom, etc.).

[0090] In the embodiment of the present invention, the deep learning model is trained based on historical data, with the input being the classroom environment information sequence (T, U, L, G) and the classroom usage status, and the output being the control strategy of the device;

[0091] Deep learning model training formula:

[0092] f(S, O) → C;

[0093] Among them, f is the deep learning model, S is the comprehensive classroom environment information, O is the classroom usage status, and C is the control strategy of the device.

[0094] Among them, the control strategy is mainly the regulation suggestions for classroom equipment (such as air conditioners, lighting, audio systems, etc.).

[0095] Furthermore, the training process of the deep learning model includes the following steps:

[0096] Data preprocessing: The collected environmental information (T, U, L, G) needs to be normalized or standardized to ensure the consistency and stability when the data is input into the deep learning model.

[0097] Model architecture design: The deep learning model usually adopts a multi-layer neural network architecture, which may include fully connected layers, convolutional layers, and recurrent layers, etc., and is selected according to the nature of the data and the complexity of the problem. This model learns the mapping relationship between the classroom environment and equipment control through historical data.

[0098] In a possible embodiment, the core of the deep learning model is to be able to extract patterns from the input environmental information sequence (T, U, L, G) and output reasonable equipment control strategies. Different deep learning architectures can be selected according to the data type and requirements. Specifically, the input layer receives multiple environmental features (T, U, L, G), where each feature can be a time series (such as temperature data in the past few hours), or a single time point data. Furthermore, due to the time dependence of the classroom environment, it is suitable to use a recurrent neural network (RNN) or a long short-term memory network (LSTM) to process these time series data. These networks can learn the dynamic changes of the input data in the time dimension. LSTM is particularly suitable for long-term dependencies and can effectively capture the change patterns of environmental features such as temperature and light over time. Further, if spatial data (such as image data of seat occupancy) needs to be analyzed, a CNN can be used to extract spatial features. After the time series information is extracted, a fully connected layer is used to further process these features to generate specific equipment control strategies. The output layer can adopt a regression model (such as outputting continuous values, such as the temperature adjustment amplitude) or a classification model (such as judging whether the equipment is turned on, adjusted, etc.) according to the requirements of the model.

[0099] Training process: Use historical data to train the model. The input includes environmental information (T, U, L, G) of multiple time steps, and the output is the control strategy corresponding to the time steps. These control strategies may include the on / off operation of the air conditioner, the temperature adjustment amplitude, the light intensity adjustment, etc. Use the data in the training set to iteratively train the deep learning model. In each round of training, the model adjusts the network weights through backpropagation to minimize the loss function. During the training process, the early stopping method can be used to prevent overfitting, that is, if the loss of the validation set does not improve for several consecutive rounds, the training is stopped.

[0100] Optimization and Evaluation: The model is trained using backpropagation algorithm and optimizers (such as AdamGD) to minimize the loss function (e.g., mean squared error loss). The generalization ability of the model is evaluated through means such as cross-validation to ensure good prediction performance in practical applications.

[0101] In the embodiments of the present invention, the control strategy optimization algorithm is based on the output of the deep learning model, combines the energy consumption model and the security model, optimizes the control strategy of the device, and generates an optimized control strategy.

[0102] Energy Consumption Model: Based on the historical classroom environment information sequence, the actual control strategy of the device, the course scheduling data of the educational administration system, and the real-time data of the campus security monitoring system, an energy consumption dataset is generated.

[0103] An energy consumption model is constructed through the energy consumption dataset to predict the energy consumption of the device.

[0104] Security Model: Based on the historical classroom environment information sequence, the actual control strategy of the device, the course scheduling data of the educational administration system, and the real-time data of the campus security monitoring system, a security dataset is generated.

[0105] A security model is constructed through the security dataset to evaluate the security of the device control strategy.

[0106] Control Strategy Optimization Algorithm Formula:

[0107] C final = argmin(λ1·E + λ2·S);

[0108] where C final represents the optimized control strategy, E represents energy consumption, S represents security, and λ1 and λ2 are the weight coefficients of energy consumption and security respectively.

[0109] In a possible implementation, the construction of the energy consumption model and the security model depends on multiple data sources, including the historical classroom environment information sequence, the actual control strategy of the device, the course scheduling data of the educational administration system, and the real-time data of the campus security monitoring system. These data work together to provide a comprehensive view of the operation of the devices in the classroom.

[0110] The energy consumption E and security S need to be normalized or standardized to ensure the consistency and stability of the data when input into the deep learning model.

[0111] Specifically, by analyzing historical data and real-time data, the energy consumption model aims to predict the energy consumption of devices. For example, changes in temperature and humidity in the classroom environment, the on / off state of devices, and their operating time may all affect energy consumption. The class schedule data can provide the usage frequency and time of classrooms, thus helping to infer the energy consumption requirements of devices during different time periods. This information is trained by a deep learning model to generate an energy consumption prediction model that can dynamically predict the energy consumption of devices during actual use.

[0112] Furthermore, the security model constructs an evaluation model for the safety of device operations by collecting monitoring data inside and outside the classroom, device control information, and class schedule data. It can evaluate whether there are potential safety hazards in the control strategy of devices. For example, if a device remains on for a long time when the classroom is unoccupied, there may be electrical faults or fire hazards. This security model can identify and mark these potential safety risks, providing a basis for optimizing the control strategy.

[0113] The core of optimizing the control strategy is through the optimization formula: C final = argmin(λ1·E + λ2·S);

[0114] where C final represents the optimized control strategy, E represents energy consumption, S represents safety, and λ1 and λ2 are the weight coefficients of energy consumption and safety respectively. The optimization goal of this formula is to simultaneously minimize the energy consumption of the device and ensure the safety of device operation.

[0115] During the optimization process, the deep learning model combines energy consumption prediction and safety assessment, inputs historical and real-time data into the model, and continuously adjusts control parameters such as the on / off state and operating time of devices through algorithms, so that the finally generated control strategy can find the best balance between energy conservation and safety. There may be a trade-off between energy consumption and safety, so the weights of λ1 and λ2 need to be adjusted according to actual needs during the optimization process. For example, in some cases, more emphasis may be placed on safety (λ2 is larger), while in other cases, reducing energy consumption may be given priority (λ1 is larger).

[0116] This intelligent control system, through the combination of deep learning with energy consumption and security models, not only realizes the efficient energy consumption management of devices in the classroom but also improves the safety of device operation, and has strong practical application value.

[0117] In the embodiment of the present invention, the energy consumption model adopts a linear regression algorithm, and the expression of the linear regression algorithm is:

[0118] E = β0 + β1T + β2U + β3L + β4G + β5C + β6S c + β7S s + ∈;

[0119] Among them, E represents the energy consumption of the device, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, C is the control strategy of the device, and S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system. β0 represents the intercept term of the model, β1 to β7 are the weights of the model, and ∈ represents the error term, that is, the difference between the actual energy consumption and the predicted energy consumption of the model. The weights of the model are optimized by the least squares method.

[0120] In a possible implementation, first, the system needs to collect data from multiple sensors in the classroom. These sensors respectively monitor the environmental temperature, object movement, light intensity, and seat occupancy. These data are uploaded to the control system in real time through the sensors or the monitoring system. Further, once sufficient historical data is collected, the least squares method is used to train the linear regression algorithm to obtain the optimal weight coefficients. The least squares method is a standard regression method that optimizes the weight coefficients by minimizing the sum of the squares of the errors between the actual observed values and the predicted values, so that the model can accurately predict the energy consumption of the device under various environmental conditions. Further, after the model training is completed, the system calculates the energy consumption of the device through the real-time input temperature, object movement, light intensity, and seat occupancy information. Using the expression of the regression model, the weighted calculation of each parameter value is performed to obtain the current energy consumption prediction value of the device. This energy consumption prediction value will be used for the decision-making of the control system to indicate whether the device should be turned on or off or adjust its operating mode (such as air conditioner, light, etc.).

[0121] The energy consumption prediction method based on the linear regression algorithm not only improves the energy efficiency of the device through accurate parameter optimization and real-time data processing, but also can intelligently adjust the device operation according to the specific usage situation of the classroom, significantly reduce energy waste, and enhance the adaptability and intelligence level of the system.

[0122] In the embodiment of the present invention, the security model adopts a logistic regression model, and the expression of the logistic regression model is:

[0123]

[0124] Among them, P represents the security probability of the device control strategy, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, C is the control strategy of the device, and S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system. α0 to α7 are the weights of the model, and the weights of the model are optimized by the gradient descent method.

[0125] In a possible implementation, a logistic regression model is used to predict the security of a device control strategy. The model calculates the security probability of device operation by inputting environmental parameters (temperature, object movement, light intensity, and seat occupancy) in the classroom. This probability value reflects whether the device control strategy is secure under a given environment, with a value range between 0 and 1. The closer it is to 1, the safer it is, and the closer it is to 0, the lower the security level.

[0126] Furthermore, the system collects real-time data through sensors in the classroom. The data content includes heat source information, object movement information, light intensity, and seat occupancy. These data are regularly uploaded to the central control system for processing and analysis.

[0127] Furthermore, to ensure that the logistic regression model can accurately predict the security of device operation, historical data needs to be analyzed first. The collected environmental parameters and the security labels of the device control strategy are used as the training set. Then, the gradient descent method is used to optimize the weight parameters of the model. Among them, the gradient descent method is an optimization algorithm that continuously adjusts the weight parameters to make the output probability value of the model as close as possible to the actual security label. Through continuous iterative optimization, the model can find the optimal weights so that under given environmental conditions, the predicted security value can reflect the true security of the device control strategy.

[0128] Once the model is trained and the weight parameters are optimized, the system can use this model to monitor the classroom environment in real time. Whenever the sensors upload new data, the model calculates the security probability of the device control strategy based on the current environmental parameter values. If this security probability is lower than a certain threshold, the system will take corresponding measures, such as reminding the administrator to check the device or automatically stopping the device operation, thus avoiding potential security problems.

[0129] In the embodiment of the present invention, the execution unit receives a control command from the device control strategy generation module. The control command includes the on / off state and adjustment parameters of the device;

[0130] Execution unit control command generation formula:

[0131] C unit =f(C final ) where C unit represents the specific control signal generated by the execution unit, C final represents the optimized control strategy, and f represents any function that converts C final to C unit .

[0132] In a possible implementation, first, the system collects various environmental data (such as temperature, light intensity, seat occupancy, etc.) in the classroom in real time through sensors and the environmental monitoring module, and uses pre-trained intelligent algorithms (such as logistic regression, machine learning, etc.) to analyze this data and generate appropriate control commands. These control commands will include the on / off state of the device and adjustment parameters (such as temperature setting, light brightness, etc.).

[0133] Furthermore, after receiving the control commands from the device control strategy generation module, the execution unit will process these commands to generate corresponding specific control signals. The core function of the execution unit is to ensure that the device can operate precisely according to the instructions. Specifically, the execution unit converts the final control command into an actual control signal according to the formula C

[0134] C unit =f(C final ). This control signal is a signal that can be recognized and executed at the device hardware level. For example, C final may indicate that the device switch state is "on" and the temperature adjustment parameter is 22°C. After receiving this command, the execution unit generates a specific control signal containing these parameters through a function, instructing the temperature control device to start working and adjust the temperature to 22°C.

[0135] Furthermore, the control signal C unit generated by the execution unit will be transmitted to the corresponding device (such as air conditioner, lighting system, etc.), and the device will perform corresponding operations according to the received signal, such as adjusting the temperature, adjusting the light, etc.

[0136] In the embodiment of the present invention, the central management unit synchronizes data with the educational administration system and the campus security monitoring system through the wireless communication module to obtain the latest course scheduling data and monitoring data;

[0137] Central management unit data synchronization formula:

[0138] D sync =f(S c , S s );

[0139] where D sync represents the synchronized data, S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system.

[0140] In a possible implementation, the educational administration system and the campus security monitoring system send real-time data to the central management unit through a wireless communication module. Further, after receiving the data, the central management unit uses a synchronization algorithm to combine the class scheduling data of the educational administration system and the security data of the security monitoring system to generate a comprehensive data set. This data set contains information such as the usage of classrooms, teaching activity arrangements, and security monitoring status. Further, based on the synchronized data, the central management unit can intelligently adjust classroom equipment. For example, if the system detects that a lecture is being held in a certain classroom and there is a security risk in that area (such as a camera detecting suspicious behavior), the central management unit can immediately notify the security system to strengthen the monitoring of that area, or send an alarm message to the classroom, requiring relevant personnel to strengthen security precautions. Further, the central management unit continuously monitors the data streams of the educational administration system and the security monitoring system to ensure that the equipment and security status always meet the current environmental requirements. When the class schedule changes or the security status is abnormal, the system can quickly respond and adjust the equipment control strategy.

[0141] In this intelligent control system, the central management unit plays a core role in coordination and control. It performs real-time data synchronization with the educational administration system and the campus security monitoring system through a wireless communication module (such as Wi-Fi, Bluetooth, Zigbee, etc.). The specific process is as follows:

[0142] The educational administration system (usually including class scheduling information, course arrangements, teacher information, etc.) sends the class scheduling data of classrooms to the central management unit regularly or in real-time through a wireless communication module. The class scheduling data S c mainly includes the usage time, teaching content, teacher information, etc. of each classroom. After obtaining this information, the central management unit can automatically adjust the working status of classroom equipment according to the course arrangement. For example, it can automatically adjust the classroom temperature, lighting, and seat status according to the requirements of different courses.

[0143] Further, the security monitoring system provides real-time monitoring data of various areas on campus (such as camera images, access control system information, alarm data, etc.). These data S s will be transmitted to the central management unit through a wireless communication module. By analyzing these data, the central management unit can detect the security situation in the classroom or the surrounding area. For example, if a suspicious person approaches the classroom or an abnormal event occurs, the central management unit can immediately adjust the security equipment status of the classroom (such as access control, alarm system, etc.) to improve the security of the classroom.

[0144] Further, the core task of the central management unit is to synchronize the data information from the educational administration system and the security monitoring system to generate a unified data set D sync . This synchronized data set includes the latest class scheduling information and security monitoring data.

[0145] By synchronizing the data of the educational administration system and the security monitoring system, the central management unit can achieve more intelligent and precise device control, which not only improves the environmental quality and security of the classroom, but also makes the whole system more flexible, automated and efficient.

[0146] The optional implementation manners of the embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the embodiments of the present invention are not limited to the specific details in the above implementation manners. Within the technical concept scope of the embodiments of the present invention, various simple modifications can be made to the technical solutions of the embodiments of the present invention, and these simple modifications all fall within the protection scope of the embodiments of the present invention.

[0147] In addition, it should be noted that, among the various specific technical features described in the above specific implementation manners, they can be combined in any suitable manner without contradiction. To avoid unnecessary repetition, the embodiments of the present invention will not separately describe various possible combination manners.

[0148] Those skilled in the art can understand that all or part of the steps in implementing the methods of the above embodiments can be completed by instructing relevant hardware through a program. This program is stored in a storage medium, including several instructions for causing a single-chip microcomputer, a chip or a processor to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks or optical disks and other various media that can store program codes.

[0149] In addition, any combination can be made between various different implementation manners of the embodiments of the present invention, as long as it does not violate the idea of the embodiments of the present invention, and it should also be regarded as the content disclosed by the embodiments of the present invention.

Claims

1. An intelligent control system for classroom equipment, characterized in that, The system includes: A status determination module: including a multi-modal perception device, a data fusion unit, and a status determination algorithm; The data fusion unit fuses the data collected by the multi-modal perception device to generate a classroom environment information sequence; The status determination algorithm: based on the classroom environment information sequence, combined with the course scheduling data of the academic affairs system and the real-time data of the campus security monitoring system, determines the classroom usage status; A device control strategy generation module based on deep learning: including a deep learning model and a control strategy optimization algorithm; The deep learning model: trained based on historical data, with the input being the classroom environment information sequence and the classroom usage status, and the output being the control strategy of the device; The control strategy optimization algorithm: based on the output of the deep learning model, combined with the energy consumption model and the security model, optimizes the control strategy of the device to generate an optimized control strategy; A distributed execution module: including multiple execution units and a communication network; The execution unit: each execution unit corresponds to a device in a classroom. After receiving a control command, the execution unit interacts through a wireless communication module to achieve the switching and adjustment of the device; The wireless communication module: adopts a power line-based half-duplex communication protocol to support the collaborative control between devices; The central management unit conducts data interaction with the status determination module, the device control strategy generation module, and the distributed execution module through the wireless communication module. The central management unit also synchronizes data with the academic affairs system and the campus security monitoring system.

2. The intelligent control system for classroom equipment according to claim 1, wherein The multi-modal perception device includes infrared sensors, ultrasonic sensors, light intensity sensors, and weight sensors distributed in the classroom; Among them, the infrared sensors: installed at the four corners and the door of the classroom, used to detect the heat source distribution in the classroom; The ultrasonic sensors: installed on the surrounding walls of the classroom, used to detect the movement of objects in the classroom; The light intensity sensors: installed on the ceiling of the classroom, used to detect the light intensity in the classroom; The weight sensors: installed under each seat in the classroom, used to detect the occupancy of the seats.

3. The intelligent control system for classroom equipment according to claim 2, wherein, The data fusion unit fuses the data collected by the multi-modal perception device to generate a classroom environment information sequence (T, U, L, G); Data fusion formula: S = αT + βU + γL + δG; Among them, S represents the comprehensive classroom environment information, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, and α, β, γ, δ are the weight coefficients of each piece of information respectively.

4. An intelligent control system for classroom equipment according to claim 3, characterized in that, The status determination algorithm is based on the classroom environment information sequence (T, U, L, G), combined with the course scheduling data of the academic affairs system and the real-time data of the campus security monitoring system, and determines the classroom usage status through a preset decision tree model; The decision tree model: generates a decision tree model based on the heat source information, object movement information, light intensity information, and seat occupancy information of the classroom, as well as the course scheduling data of the academic affairs system and the real-time data of the campus security monitoring system; The training dataset of the decision tree model includes the historical classroom environment information sequence and the corresponding classroom usage status. By training, the parameters of the model are optimized so that the model can accurately determine the classroom usage status.

5. The intelligent control system for classroom equipment according to claim 1, wherein The deep learning model is trained based on historical data. The input is the classroom environment information sequence (T, U, L, G) and the classroom usage status, and the output is the control strategy of the device; Training formula of the deep learning model: f(S, O) → C; where f is the deep learning model, S is the comprehensive classroom environment information, O is the classroom usage status, and C is the control strategy of the device.

6. The intelligent control system for classroom equipment according to claim 1, wherein, Based on the output of the deep learning model, the control strategy optimization algorithm combines the energy consumption model and the security model to optimize the control strategy of the device and generate an optimized control strategy; Energy consumption model: Based on the historical classroom environment information sequence, the actual control strategy of the device, the course scheduling data of the academic affairs system, and the real-time data of the campus security monitoring system, an energy consumption data set is generated; An energy consumption model is constructed through the energy consumption data set to predict the energy consumption of the device; Security model: Based on the historical classroom environment information sequence, the actual control strategy of the device, the course scheduling data of the academic affairs system, and the real-time data of the campus security monitoring system, a security data set is generated; A security model is constructed through the security data set to evaluate the security of the device control strategy; Control strategy optimization algorithm formula: C final = argmin(λ1·E + λ2·S); Among them, C final represents the optimized control strategy, E represents the energy consumption, S represents the safety, and λ1 and λ2 are the weight coefficients of the energy consumption and the safety respectively.

7. The intelligent control system for classroom equipment according to claim 6, wherein, The energy consumption model uses a linear regression algorithm, and the expression of the linear regression algorithm is: E = β0 + β1T + β2U + β3L + β4G + β5C + β6S c + β7S s + ∈; Among them, E represents the energy consumption of the device, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, C is the control strategy of the device, and S c represents the course scheduling data of the academic affairs system, and S s represents the real-time data of the campus security monitoring system. β0 represents the intercept term of the model, β1 to β7 are the weights of the model, and ∈ represents the error term, that is, the difference between the actual energy consumption and the predicted energy consumption of the model. The weights of the model are optimized by the least squares method.

8. The intelligent control system for classroom equipment according to claim 7, characterized in that, The security model uses a logistic regression model, and the expression of the logistic regression model is: Among them, P represents the security probability of the device control strategy, T represents the heat source information, U represents the object movement information, L represents the light intensity information, G represents the seat occupancy information, C is the control strategy of the device, and S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system. α0 to α7 are the weights of the model, and the weights of the model are optimized by the gradient descent method.

9. The intelligent control system for classroom equipment according to claim 1, wherein The execution unit receives control commands from the device control strategy generation module, and the control commands include the on / off state and adjustment parameters of the device; Execution unit control command generation formula: C unit = f(C final ) where C unit represents the specific control signal generated by the execution unit, and C final represents the optimized control strategy. f represents any function that converts C final to C unit .

10. The intelligent control system for classroom equipment according to claim 1, characterized in that, The central management unit synchronizes data with the academic affairs system and the campus security monitoring system through the wireless communication module to obtain the latest course scheduling data and monitoring data; Central management unit data synchronization formula: D sync = f(S c , S s ); Among them, D sync represents the data after synchronization, and S c represents the course scheduling data of the educational administration system, and S s represents the real-time data of the campus security monitoring system.

Images