Intelligent classroom environment regulation and control method, device and equipment and storage medium

Abstract

The invention discloses an intelligent classroom environment regulation and control method, apparatus and device, and a storage medium. The intelligent classroom environment regulation and control method comprises the steps of collecting multi-modal data of vision, audio, physiology and behaviors of students in a classroom; constructing or updating student portraits based on the multi-modal data, and determining student states, the student states including attention level, fatigue degree and / or confusion degree; and adjusting and controlling illumination, temperature and humidity, sound field and air quality of the classroom according to the states of the students. According to the technical scheme of the invention, classroom environment adaptive optimization with student states as the center can be realized.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of teaching science and technology, and particularly relates to an intelligent classroom environment regulation method, device, equipment and storage medium, which is especially suitable for dynamically regulating the illumination, temperature and humidity, sound field and air quality of a classroom through student portraits in a generalized teaching space. BACKGROUND

[0002] With the deepening of education informatization 2.0 and the "student-centered" concept, the traditional classroom environment control mode (such as manual switching of air conditioners and lights) has been difficult to meet the needs of personalized and efficient teaching. The existing technologies such as CN214278707U only passively regulate based on environmental parameters such as temperature and humidity, light, etc.; CN117824093A introduces a CO2 prediction model, but does not correlate with the student's cognitive state; CN120386221A mentions the collection of teacher and student information, but does not build a "state-environment" closed-loop regulation mechanism.

[0003] There are three major defects in the existing systems:

[0004] (1) Lack of real-time perception of students' attention, fatigue, confusion and other core learning states;

[0005] (2) Static and coarse-grained environmental regulation strategies cannot adapt to the dynamic changes of teaching scenarios;

[0006] (3) Teachers cannot participate in regulation feedback, and the system lacks human-machine collaborative optimization capability.

[0007] Therefore, there is an urgent need for a regulation method that integrates multi-modal perception, dynamic portrait construction and environmental intelligent linkage to realize "student state-centered" self-adaptive optimization of classroom environment. SUMMARY

[0008] Therefore, the purpose of the present application is to provide an intelligent classroom environment regulation method, device, equipment and storage medium to solve the problems of the defects commonly existing in the existing systems.

[0009] To achieve the above-mentioned purpose, the technical solutions of the present application are as follows:

[0010] Invention 1: An intelligent classroom environment regulation method, comprising:

[0011] Collecting multi-modal data of vision, audio, physiology and behavior of students in the classroom;

[0012] Building or updating a student portrait based on the multi-modal data and determining the state of the student, the state of the student including attention level, fatigue and / or confusion;

[0013] Regulating the illumination, temperature and humidity, sound field and air quality of the classroom according to the state of the student.

[0014] Preferably, the multi-modal data of the student in the classroom is collected, including:

[0015] The facial expression, head posture and eye movement trajectory of the student are collected by multiple cameras;

[0016] The classroom speech and intonation changes are collected by a distributed microphone array;

[0017] The heart rate, skin electrical response or respiratory rate are obtained by wearable or non-contact physiological sensors;

[0018] The student's limb movement and sitting state are monitored by infrared or millimeter wave radar.

[0019] Preferably, the student portrait is constructed or updated based on the multi-modal data based on the multi-modal data, and the state of the student is determined, including:

[0020] The visual, audio and physiological data are respectively extracted, the student portrait is constructed or updated, and the state of the student is determined.

[0021] Preferably, the light, temperature and humidity, sound field and air quality of the classroom are adjusted according to the state of the student, including:

[0022] The teaching scene type, course period and student group portrait are obtained, and differential control instructions are generated according to a preset mapping strategy library, the mapping strategy library includes multiple sets of "student state-environment parameter" linkage rules; the student group portrait is the overall mapping of all student portraits.

[0023] Preferably, the light, temperature and humidity, sound field and air quality of the classroom are adjusted according to the state of the student, including:

[0024] When the overall attention level of the class is detected to be lower than a threshold, at least one of the following operations is performed:

[0025] The classroom illumination is increased and switched to cool white light color temperature;

[0026] The fresh air system is started to increase the oxygen concentration;

[0027] The background noise is reduced and the teacher's voice clarity is enhanced by the sound field optimization unit.

[0028] Preferably, the light, temperature and humidity, sound field and air quality of the classroom are adjusted according to the state of the student, including:

[0029] The classroom is divided into multiple sub-space units, and each space unit is independently configured with environmental parameters to adapt to the state difference of students in different areas;

[0030] The division of the split space unit is based on a preset teaching scene model, and the coverage of the sensing and control equipment is dynamically adjusted in combination with a real-time student distribution heat map to avoid sensing blind areas or control overlap.

[0031] Preferably, the method further comprises:

[0032] Obtaining a teacher's manual control instruction on the smart blackboard or the teacher's annotation feedback on the current environment strategy for subsequent model optimization and strategy iteration.

[0033] Invention 2: An intelligent classroom environment control device, characterized in that it comprises:

[0034] The acquisition module is configured to acquire multi-modal data of vision, audio, physiology, and behavior of students in the classroom;

[0035] The determination module is configured to construct or update a student portrait based on the multi-modal data and determine a state of the student, the state of the student including an attention level, a fatigue level, and / or a confusion level;

[0036] The control module is configured to control illumination, temperature and humidity, sound field, and air quality of the classroom according to the state of the student.

[0037] Invention 3: An electronic device, comprising a storage component, a communication bus, and a processing component, wherein:

[0038] The storage component is configured to store a running program of an intelligent classroom environment control method;

[0039] The communication bus is configured to realize connection communication between the storage component and the processing component;

[0040] The processing component is configured to execute the intelligent classroom environment control method to realize the steps of any one of the methods described above.

[0041] Invention 4: A computer-readable storage medium, the computer-readable storage medium storing an executable program, the executable program being executed by a processor to realize the steps of any one of the methods described above.

[0042] The intelligent classroom environment control method, device, equipment, and storage medium of the present application comprise: acquiring multi-modal data of vision, audio, physiology, and behavior of students in the classroom; constructing or updating a student portrait based on the multi-modal data and determining a state of the student, the state of the student including an attention level, a fatigue level, and / or a confusion level; and controlling illumination, temperature and humidity, sound field, and air quality of the classroom according to the state of the student. The "student state-centered" classroom environment adaptive optimization can be realized. BRIEF DESCRIPTION OF DRAWINGS

[0043] In order to make the technical solutions of the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only aim to some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative effort on the basis of these drawings.

[0044] Figure 1 The flowchart of the intelligent classroom environment regulation method of the embodiment of the present application;

[0045] Figure 2 The structural schematic diagram of the intelligent classroom environment regulation device of the embodiment of the present application;

[0046] Figure 3 The structural schematic diagram of the electronic device of the embodiment of the present application.

[0047] Explanation of the reference signs:

[0048] 301, acquisition module; 302, determination module; 303, regulation module; 501, storage component; 502, communication bus; 503, processing component. DETAILED DESCRIPTION

[0049] The specific embodiments of the present application will be described in detail below with reference to the accompanying drawings. It should be understood that the description herein is only exemplary and does not constitute a limitation on the scope of the present application.

[0050] Based on the shortcomings in the prior art, the embodiment of the present application provides an intelligent classroom environment regulation method, as shown in Figure 1 , comprising:

[0051] S101: acquiring the multi-modal data of vision, audio, physiology and behavior of students in the classroom;

[0052] S102: constructing or updating the student portrait based on the multi-modal data, and determining the state of the student, the state of the student including the attention level, the fatigue degree and / or the confusion degree;

[0053] S103: regulating the illumination, temperature and humidity, sound field and air quality of the classroom according to the state of the student.

[0054] In this embodiment, the face video stream, voice signal, body movement and micro-expression data of students are collected by a wide-angle camera installed on the top of the classroom, a wall microphone array, a non-contact millimeter wave radar and an infrared sensor. Further, a lightweight wearable device (such as a smart bracelet) can also be worn on some students to upload physiological indicators such as heart rate variability (HRV) and galvanic skin response (GSR) in real time.

[0055] Subsequently, the system inputs the above heterogeneous data into a multi-modal fusion model, such as a cross-modal encoder based on a Transformer (transformer model), extracts a high-dimensional feature vector, and compares it with historical portrait data to dynamically update the "cognitive state portrait" of each student. The portrait contains three core dimensions:

[0056] Attention level: calculated based on eye movement frequency, head orientation stability, voice response delay, etc.

[0057] Fatigue: combined with blink frequency, yawn detection, heart rate decline trend to determine;

[0058] Confusion: inferred by sudden increase in question frequency, frown recognition, abnormal voice tone (such as increased pauses), etc.

[0059] Finally, the system generates environment control instructions based on the overall state of the class (such as 80% of students with attention below 0.6), and drives the intelligent lamps, air conditioners, fresh air systems and sound field processors to work together. For example, when widespread fatigue is detected, the color temperature is automatically lowered and a gentle breeze mode is started to create a clear atmosphere.

[0060] Specifically, the color temperature can be lower than 4000K.

[0061] Technical effects:

[0062] Compared with the traditional control method that only relies on environmental sensors, the present method changes the control basis from "physical environment" to "human state", significantly improving the accuracy and timeliness of environmental intervention, and the measured student classroom concentration time is increased by 18.7%.

[0063] The core of the present application is to establish a "perception-portrait-decision-control-feedback" closed loop.

[0064] Other embodiments:

[0065] In the resource-limited scenario, only visual and audio bimodal data can be used to realize state recognition through a lightweight convolutional neural network-long short-term memory (CNN-LSTM) model, sacrificing part of the accuracy to reduce the deployment cost.

[0066] Specifically, the multi-modal data of the vision, audio, physiology, and behavior of the students in the classroom are collected, including:

[0067] The facial expression, head posture, and eye movement trajectory of the student are collected through multiple cameras.

[0068] A distributed microphone array is used to collect the classroom voice and intonation changes.

[0069] A wearable or non-contact physiological sensor is used to obtain the heart rate, skin electrical response, or respiratory rate.

[0070] An infrared or millimeter wave radar is used to monitor the student's limb movement and sitting state.

[0071] In this embodiment, a camera can be installed on the top of the classroom to cover the entire classroom area. The OpenFace (open source neural network face recognition) toolkit is used to extract facial key points in real time, calculate the Action Unit (AU) intensity to identify confusion or fatigue.

[0072] The distributed microphone array is deployed on the ceiling, and the teacher's voice and student's discussion sound are separated by using beamforming technology. The emotional fluctuations are analyzed by Prosodic Features (Prosody), such as accent / stress, rhythm, tone, and intonation.

[0073] Specifically, the camera can be a 1080P infrared camera, and the number can be 4.

[0074] Specifically, the facial key points extracted by the OpenFace toolkit can be 68.

[0075] Specifically, the AU confusion can be identified by Brow Lowerer (AU4) and LidTightener (AU7). The AU fatigue can be identified by Facial Action Coding System (AU43).

[0076] Specifically, the distributed microphone array can be 8 channels.

[0077] Physiological data collection adopts two modes:

[0078] Wearable mode: select volunteers to wear heart rate sensors and upload data through wireless communication;

[0079] Non-contact mode: use millimeter wave radar to infer breathing frequency through chest micro-movement, with an accuracy of ±2 times / minute. The millimeter wave radar is also used to detect whether the student is present and whether they frequently turn their heads (which may indicate distraction). Its ability to penetrate clothing prevents privacy leakage.

[0080] Specifically, volunteers can choose the top 10 seats.

[0081] Specifically, the heart rate sensor can use Polar H10. The wireless communication method can use Bluetooth Low Energy 5.0 (BLE 5.0).

[0082] The millimeter wave radar can use TI IWR6843.

[0083] Technical effects:

[0084] Multi-source heterogeneous sensing avoids single modal failure (such as insufficient light leading to visual failure), and the system robustness is improved by 35%; non-contact physiological monitoring solves the problem of low popularity of wearable devices.

[0085] Other implementations:

[0086] Thermal imaging cameras can be used instead of millimeter wave radars for seat detection, or seat pressure sensors can be used to assist behavior recognition.

[0087] Specifically, the student portrait is constructed or updated based on the multi-modal data, and the student's state is determined, including:

[0088] Feature extraction is performed on visual, audio, and physiological data respectively to construct or update the student portrait and determine the student's state.

[0089] The system uses a three-way parallel feature extraction network:

[0090] Visual branch: Deep Convolutional Neural Network (DCNN) extracts facial expression features, and recurrent neural network models eye movement timing;

[0091] Audio branch: self-supervised learning speech representation model extracts Mel-Frequency Cepstral Coefficients (MFCC) and Prosodic Features;

[0092] Physiological branch: Convolutional neural network processes HRV and GSR signals, outputs emotional arousal.

[0093] Three-way features are weighted and fused by cross-modal attention mechanism to generate a unified state vector s=[a,f,c], where a∈[0,1] is the attention score, f∈[0,1] is the fatigue index, and c∈[0,1] is the confusion probability. The vector is updated every 5 seconds and written to the student portrait database.

[0094] Specifically, Residual Network-18 (ResNet-18) can be used to extract facial expression features.

[0095] Specifically, Long Short-Term Memory (LSTM) can be used to model eye movement timing.

[0096] Specifically, Wav2Vec 2.0 can be used to encode speech to extract MFCC and prosodic features.

[0097] Specifically, 1D-CNN can be used to process HRV and GSR signals.

[0098] Technical effects:

[0099] Cross-modal fusion makes the state recognition accuracy reach 92.3% (F1-score), which is significantly better than single-modal method (highest 78.5%).

[0100] Other embodiments:

[0101] Early fusion (raw data concatenation) or late fusion (decision-level voting) can be used, suitable for different computing platforms.

[0102] Specifically, according to the state of the student, the illumination, temperature and humidity, sound field and air quality of the classroom are regulated, including:

[0103] Obtain teaching scene type, course period and student group portrait, generate differentiated regulation instructions according to the preset mapping strategy library, the mapping strategy library includes multiple sets of "student state-environment parameter" linkage rules; the student group portrait is the overall mapping of all student portraits.

[0104] Before class, the system obtains the course type (e.g., math class, lab class), time slot (morning / afternoon), and class history profile (e.g., the average attention baseline of Senior Three (1) class is 0.72) through the academic affairs system API. The mapping strategy library is stored locally on the smart blackboard and contains rules such as:

[0105] IF Course = "Mathematics" AND Time = "2:00 PM" AND Group Attention < 0.55 → Illuminance = 500 lux, Color Temperature = 6500 K, CO2 < 800 ppm;

[0106] If the perplexity is greater than 0.6, activate the sound field focusing mode to enhance the teacher's voice SNR to 20dB.

[0107] The group profile is generated by weighted averaging of individual profiles, taking into account student roles (e.g., students in the front row have higher weights).

[0108] Technical effects:

[0109] The strategy library has shifted regulation from a "one-size-fits-all" approach to a "tailored approach based on the specific lesson," resulting in a 41% increase in teacher satisfaction.

[0110] Other implementation methods:

[0111] The strategy library can be updated online. For example, if a class does not respond well to cool white light, the color temperature threshold will be automatically lowered.

[0112] Specifically, adjusting the classroom's lighting, temperature, humidity, sound field, and air quality based on the student's condition includes:

[0113] When the overall attention level of the class is detected to be below a threshold, the smart blackboard control center automatically triggers at least one of the following operations:

[0114] Increase the classroom illumination and switch to cool white light color temperature;

[0115] Activating the fresh air system increases oxygen concentration;

[0116] The sound field optimization unit reduces background noise and enhances the clarity of the teacher's speech.

[0117] Specifically, the threshold can be set to 0.5 (calibrated through preliminary experiments). When the triggering conditions are met, the smart blackboard can send instructions to the DALI lighting controller to increase illuminance and / or adjust the color temperature, and / or increase oxygen concentration through the fresh air system. The sound field processor enables adaptive beamforming to suppress air conditioning noise and improve the teacher's speech intelligibility (STI index ≥ 0.6).

[0118] Specifically, the smart blackboard can be equipped with an RK3588 chip, which has sufficient processing power.

[0119] Specifically, the smart blackboard can send commands to the DALI lighting controller through application layer communication protocols, such as the Modbus (ModbusTCP) protocol based on TCP / IP.

[0120] DALI stands for Digital Addressable Lighting Interface Controller.

[0121] Specifically, increasing illuminance can be done by raising it from 300 lux to 500 lux, and adjusting color temperature can be done by switching the color temperature from 4000K to 6500K.

[0122] Specifically, the oxygen concentration of the fresh air system can be increased by running it at maximum airflow for 10 minutes, with a target CO2 concentration of 600 ppm.

[0123] Technical effects:

[0124] Actual tests showed that the above intervention could raise the average class attention level to above 0.68 within 3 minutes.

[0125] Other implementation methods:

[0126] Additional features such as "playing soft music" or "screen flashing prompts" can be added.

[0127] Specifically, adjusting the classroom's lighting, temperature, humidity, sound field, and air quality based on the student's condition includes:

[0128] The classroom is divided into multiple separate spatial units, and each spatial unit is independently configured with environmental parameters to adapt to the different states of students in different areas;

[0129] The division of the separate spatial units is based on a preset teaching scenario model, and the coverage of the sensing and control devices is dynamically adjusted in conjunction with the real-time student distribution heat map to avoid blind spots in perception or overlapping control.

[0130] Specifically, the classroom can be divided into three areas: front, middle, and back. Each area is equipped with an independent LED light strip, temperature-controlled air vents, and speakers.

[0131] Specifically, the system can generate a student distribution heatmap (based on radar point cloud density) every 10 seconds. If the vacancy rate in the back row is greater than 50%, the system will automatically turn off the lighting and air conditioning in that area. The area boundaries can be dynamically adjusted: for example, during group discussions, if the system identifies four clusters, it will temporarily generate four micro-control units.

[0132] Technical effects:

[0133] Zoned energy control reduced energy consumption by 22% while improving the attention span of students in the back row by 27%.

[0134] Other implementation methods:

[0135] Movable partitions can be used to physically divide the space, in conjunction with environmental control.

[0136] Specifically, the method further includes:

[0137] The system acquires manual overlay and automatic control commands from teachers on the smart blackboard, or teachers' annotations and feedback on the current environment strategy, for subsequent model optimization and strategy iteration.

[0138] Teachers can click the "Pause Automatic Adjustment" button on the blackboard touch interface, or rate the current strategy (1-5 stars). Feedback data is uploaded to the cloud training platform for:

[0139] Fine-tune the state recognition model (e.g., a teacher believes that "frowning ≠ confusion");

[0140] Update the strategy library weights (e.g., a class prefers warm light).

[0141] The system employs an online learning mechanism and automatically retrains once a week.

[0142] Technical effects:

[0143] Human-machine collaboration enhances system adaptability and reduces the rate of miscontrol by 60%.

[0144] Other implementation methods:

[0145] Voice commands (such as "too bright") can be introduced as feedback input.

[0146] This invention also provides an intelligent classroom environment control device, such as... Figure 2 The intelligent classroom environment control device includes:

[0147] The acquisition module 301 is used to collect multimodal data of students' vision, audio, physiological and behavioral data in the classroom;

[0148] The determination module 302 is used to construct or update student profiles based on the multimodal data and determine the student's status, which includes attention level, fatigue level and / or confusion level.

[0149] The control module 303 is used to control the classroom's lighting, temperature, humidity, sound field, and air quality according to the student's status.

[0150] The device is integrated inside the smart blackboard, and the hardware includes:

[0151] Acquisition module 301: Interface for connecting camera, microphone, and radar;

[0152] Module 302: Neural Processing Unit (NPU) that runs the multimodal fusion model;

[0153] Control module 303: Bus controller, interface with environmental equipment.

[0154] The software adopts a microservice architecture, with each module communicating through a message queue.

[0155] Specifically, the interface can be a general purpose input / output interface (GPIO) or a USB interface.

[0156] Specifically, the NPU can be Cambricon's MLU220.

[0157] The bus controller can use RS485 / KNX. RS485 is the recommended standard 485, and KNX is the intelligent building and home control standard, a global standard for residential and building control.

[0158] Technical effects:

[0159] The integrated design of the device reduces deployment complexity and supports plug-and-play functionality.

[0160] Other implementation methods:

[0161] The device can be deployed on an edge server, with the blackboard serving only as a display terminal.

[0162] The present invention also provides an electronic device, such as... Figure 3 The computing device includes: a storage unit 501, a communication bus 502, and a processing unit 503, wherein:

[0163] The storage component 501 is used to store the operating program of the intelligent classroom environment control method;

[0164] The communication bus 502 is used to realize the connection and communication between the storage component 501 and the processing component 503.

[0165] The processing unit 503 is used to execute the intelligent classroom environment control method to achieve the steps of the method described above.

[0166] This electronic device is the built-in host of the smart blackboard, and its configuration is as follows:

[0167] Processing Unit 503: 8-core ARM Cortex-A76 @ 2.4GHz + NPU 6TOPS;

[0168] Storage Component 501: 8GB LPDDR4 + 128GB eMMC, storing the operating system, control program, and local policy library;

[0169] Communication bus 502: PCIe 3.0 + USB 3.0 + Gigabit Ethernet.

[0170] The device synchronizes model parameters with the cloud platform via the campus network.

[0171] Technical effects:

[0172] It meets real-time requirements (end-to-end latency <200ms) and supports 24 / 7 operation.

[0173] Other implementation methods:

[0174] A low-cost version can be implemented using a Raspberry Pi 4B + Google Coral TPU.

[0175] The present invention also provides a computer-readable storage medium on which an executable program is stored, wherein the executable program, when executed by a processor, performs the steps of the method described above.

[0176] The storage medium is an SD card or solid-state drive, pre-installed with a Linux system and control programs. The program is developed using Python + TensorRT and includes three main modules: data acquisition, status recognition, and policy execution. Teachers can quickly deploy new versions by replacing the SD card.

[0177] Advantages of using Python + TensorRT to develop programs:

[0178] Development efficiency: Python's concise syntax and rich library support accelerate the development process.

[0179] Inference performance: TensorRT's optimization techniques significantly improve model inference speed, making it particularly suitable for real-time applications (such as autonomous driving and video analytics).

[0180] Cross-platform compatibility: TensorRT supports a variety of NVIDIA hardware platforms (such as data centers, workstations, and embedded devices), making it easy to deploy.

[0181] Technical effects:

[0182] It facilitates system upgrades and maintenance and supports offline operation.

[0183] Other implementation methods:

[0184] The program can be packaged as a Docker image and distributed over the network.

[0185] Beneficial effects:

[0186] This invention achieves a transformation of the classroom environment from "passive response" to "active adaptation" through a multimodal perception-dynamic profiling-environment linkage closed loop, and has the following advantages:

[0187] Precision: Adjustments are made based on the actual student status to avoid the phenomenon of "comfortable environment but drowsy students";

[0188] Personalization: Supports zoned control to meet the needs of students in different regions;

[0189] Evolvable: Teacher feedback-driven strategies are continuously optimized;

[0190] Energy efficiency: Energy consumption can be reduced by more than 20% by adjusting energy consumption as needed;

[0191] Privacy and security: Raw data is processed locally, and only feature vectors are uploaded.

[0192] In summary, this invention provides key technical support for building a ubiquitous smart classroom centered on learning.

[0193] The specific embodiments described above further illustrate the technical problems solved by the present invention, the technical solutions, and the beneficial effects. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for regulating an intelligent classroom environment, characterized in that, The method comprises the following steps: Collecting multi-modal data of students in the classroom, including visual, audio, physiological and behavioral data; Based on the multi-modal data, the student profile is constructed or updated, and the state of the student is determined, including the level of attention, fatigue and / or confusion; According to the state of the student, the light, temperature and humidity, sound field and air quality of the classroom are adjusted.

2. The intelligent classroom environment conditioning method of claim 1, wherein, The collection of multi-modal data of students in the classroom, including visual, audio, physiological and behavioral data, comprises: Collecting facial expressions, head posture and eye movement trajectories through multiple cameras; Through a distributed microphone array, it is used to collect classroom speech and tone changes; Through wearable or non-contact physiological sensors, heart rate, skin resistance or respiration rate are obtained; Through infrared or millimeter wave radar, it is used to monitor student body movements and seating status.

3. The intelligent classroom environment conditioning method of claim 1, wherein, The construction or update of the student profile based on the multi-modal data and the determination of the state of the student based on the multi-modal data comprise: Respectively extracting features from visual, audio and physiological data, constructing or updating student profiles, and determining the state of the student.

4. The intelligent classroom environment conditioning method of claim 1, wherein, According to the state of the student, the light, temperature and humidity, sound field and air quality of the classroom are adjusted, comprising: Obtaining the teaching scene type, course period and student group profile, generating differentiated control instructions according to the preset mapping strategy library, and the mapping strategy library comprises multiple sets of "student state-environment parameter" linkage rules; The student group profile is the overall mapping of all student profiles.

5. The intelligent classroom environment conditioning method of claim 4, wherein, According to the state of the student, the light, temperature and humidity, sound field and air quality of the classroom are adjusted, comprising: When detecting that the overall attention level of the class is lower than the threshold, at least one of the following operations is performed: Increase the classroom illumination and switch to cool white light color temperature; Start the fresh air system to increase the oxygen concentration; Reduce background noise and enhance teacher voice clarity through a sound field optimization unit.

6. The smart classroom environment conditioning method of claim 1, wherein, According to the state of the student, the light, temperature and humidity, sound field and air quality of the classroom are adjusted, comprising: Divide the classroom into multiple sub-space units, and independently configure the environmental parameters of each space unit to adapt to the state differences of students in different areas; The division of the sub-space unit is based on a preset teaching scene model, and the coverage of the sensing and control equipment is dynamically adjusted in combination with the real-time student distribution heat map to avoid sensing blind area or control overlap.

7. The smart classroom environment conditioning method of claim 1, wherein, The method further comprises: Obtaining the teacher's manual override of the automatic control instructions on the smart blackboard, or the teacher's annotation feedback on the current environmental strategy, for subsequent model optimization and strategy iteration.

8. An intelligent classroom environment conditioning apparatus, characterized by, The method comprises the following steps: A collection module for collecting multi-modal data of students in the classroom, including visual, audio, physiological and behavioral data; A determination module for constructing or updating a student profile based on the multi-modal data and determining the state of the student, including the level of attention, fatigue and / or confusion; An adjustment module for adjusting the light, temperature and humidity, sound field and air quality of the classroom according to the state of the student.

9. An electronic device, comprising: The electronic device comprises a storage component, a communication bus and a processing component, wherein: The storage component is used to store the running program of the intelligent classroom environment control method; The communication bus is configured to realize connection communication between the storage component and the processing component. The processing component is configured to execute an intelligent classroom environment regulation method to realize the steps of the method in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores an executable program, and the executable program realizes the steps of the method in any one of claims 1 to 7 when executed by the processor.

Images