Intelligent classroom health lighting system based on multi-sensor fusion

Abstract

This invention relates to the field of smart lighting technology, specifically to a smart classroom healthy lighting system based on multi-sensor fusion. It includes: a multi-source sensor synchronous acquisition unit; a health risk analysis unit; an adaptive fusion processing unit; a healthy lighting compliance decision-making unit; and a smart lighting adaptive drive control unit. This invention improves the assessment dimensions of lighting health risks by employing techniques such as multi-source sensor data temporal registration and coarse noise reduction preprocessing, and coupled calculation of instantaneous health risks and blue light / glare time-accumulated risks. It considers both instantaneous conditions and the cumulative impact of long-term teaching scenarios, enhancing the comprehensiveness of health risk assessment. Through techniques such as health threshold neighborhood sensitivity adjustment, heterogeneous data priority fusion, and differentiated weight control, the system can adjust indicator control weights and response strategies according to dynamic changes in the lighting environment, alleviating the problem of adjustment lag under different teaching conditions.

Description

Technical Field

[0001] This invention relates to the field of smart lighting technology, and more specifically, to a smart classroom health lighting system based on multi-sensor fusion. Background Technology

[0002] As educational settings demand increasingly higher standards for lighting quality, smart classroom healthy lighting has evolved from traditional constant illuminance control to a multi-dimensional, collaborative optimization of health indicators. Current mainstream solutions primarily utilize various sensors for illuminance, blue light, and glare to collect lighting environment data and make initial adjustments to lighting parameters based on simple threshold judgments. While this has yielded some progress in basic lighting energy conservation and health protection, technological optimization remains focused on hardware configuration and has not yet achieved a systematic breakthrough in data processing and control logic. A healthy lighting system needs to comprehensively and accurately perceive the health status of the classroom lighting environment and dynamically output adapted lighting control strategies based on the teaching scenario. However, existing technologies cannot simultaneously address the comprehensiveness of assessment, the timeliness of response, and the robustness of the system in terms of in-depth data mining and dynamic scenario adaptation.

[0003] Specifically, existing solutions have two prominent limitations: First, the health indicator assessment method is singular, relying solely on instantaneous sampling data for threshold judgment, lacking a systematic consideration of the impact of the lighting environment in long-term teaching scenarios, and failing to comprehensively reflect health risks in actual use; second, the multi-indicator collaborative control strategy lacks dynamic adaptability, employing preset fixed rules, and cannot flexibly adjust the priority of indicator control according to changes in the lighting environment. This easily leads to control lag and rigid adjustment problems in different working conditions such as blackboard writing and multimedia teaching, making it difficult to guarantee the health protection effect and hindering the large-scale promotion and application of the system. In view of this, we propose a smart classroom health lighting system based on multi-sensor fusion. Summary of the Invention

[0004] The purpose of this invention is to provide a smart classroom health lighting system based on multi-sensor fusion, in order to solve the problems of existing classroom health lighting systems mentioned in the background art, which have a single health indicator assessment method, rely solely on instantaneous data to determine health status, lack systematic consideration of the impact of light environment on long-term teaching scenarios, are difficult to fully reflect the health risks in actual use, and lack dynamic adaptability and rigid control logic of multi-indicator collaborative control strategy, which cannot flexibly adjust the control priority according to the dynamic changes of teaching conditions, and are prone to control lag and insufficient health protection effect.

[0005] To address the aforementioned technical problems, the present invention aims to provide a smart classroom health lighting system based on multi-sensor fusion, comprising: A multi-source sensor synchronous acquisition unit collects various types of raw environmental data in the smart classroom, performs synchronous processing and timestamp marking, and generates a raw sensor dataset with time sequence identifiers. The health risk analysis unit receives the original sensor dataset and sequentially performs temporal registration and coarse noise reduction processing. Based on the preset health thresholds of various sensor indicators, it constructs an instantaneous health risk evaluation function, calculates the time-cumulative risk factors corresponding to blue light and glare indicators, couples the instantaneous health risk and the time-cumulative risk factors to obtain the comprehensive health risk, and completes the risk stratification of lighting indicators based on the comprehensive health risk. An adaptive fusion processing unit receives risk stratification results, carries heterogeneous data priority fusion logic, enables a health threshold neighborhood sensitivity adjustment mechanism, increases the sampling frequency and weight gain slope of high-risk indicators of blue light and glare in critical health states, applies weight gain adjustment to high-risk health indicators of blue light and glare, applies weight constraint adjustment to conventional comfort indicators of illuminance, completes data fusion through differential normalization processing that preserves the boundary characteristics of health thresholds, and outputs high-precision lighting environment feature data. The healthy lighting compliance decision unit receives high-precision lighting environment characteristic data, completes compliance verification through the built-in classroom healthy lighting threshold judgment model, and generates healthy lighting control instructions adapted to the teaching environment by combining the real-time personnel distribution and environmental conditions in the classroom. The intelligent lighting adaptive drive control unit receives healthy lighting control commands and performs high-frequency PWM stepless dimming and color adjustment operations on the classroom's whole-area lighting terminal equipment to achieve full-area dynamic adjustment of lighting illuminance, color temperature, blue light ratio, and glare suppression.

[0006] As a further improvement to this technical solution, the multi-source sensor synchronous acquisition unit includes a sensor acquisition module, a data synchronization processing module, and a time sequence marker generation module, wherein: The sensor acquisition module collects various types of raw environmental data within the smart classroom; The data synchronization processing module performs synchronization processing on multiple types of raw environmental data collected by the sensor acquisition module. The time-series marker generation module timestamps the data after it has been synchronized by the data synchronization processing module, generating a raw sensor dataset with time-series identifiers.

[0007] As a further improvement to this technical solution, the health risk analysis unit includes a preprocessing and instantaneous risk calculation module; the preprocessing and instantaneous risk calculation module is used to purify and preprocess the original sensor dataset and quantify the instantaneous health risks of various lighting indicators; the data preprocessing and instantaneous risk calculation process of the preprocessing and instantaneous risk calculation module includes the following steps: S21.1 Receive the raw sensor dataset, perform time registration processing and coarse noise reduction processing in sequence, unify the time reference of all sensor data, and remove abnormal data in the dataset that exceed the physical range. S21.2 Construct an instantaneous health risk assessment function based on preset health thresholds for various sensor indicators, and calculate and output the instantaneous health risk values ​​corresponding to various sensor indicators. .

[0008] As a further improvement to this technical solution, the health risk analysis unit also includes a time-cumulative risk calculation module; the time-cumulative risk calculation module is used to quantify the time-series cumulative risk based on instantaneous health risk for high-risk lighting indicators; the time-cumulative risk factor calculation process of the time-cumulative risk calculation module includes the following steps: S22.1 Receive the instantaneous health risk value output by the preprocessing and instantaneous risk calculation module. From this, values ​​corresponding to two categories of indicators, blue light and glare, are selected. S22.2 Activate the preset sliding time window statistical mode, combining the time series data within the window with the instantaneous health risk value. Calculate and output the time-cumulative risk factor of the corresponding indicator. .

[0009] As a further improvement to this technical solution, the health risk analysis unit also includes a comprehensive risk coupling and stratification module; the comprehensive risk coupling and stratification module is used to integrate instantaneous risk and cumulative risk, and to complete the classification of lighting health risk levels; the risk coupling and indicator stratification process of the comprehensive risk coupling and stratification module includes the following steps: S23.1 Receive instantaneous health risk value With time-cumulative risk factors The comprehensive health risk value of each lighting indicator is obtained by coupling through a weighted fusion method. ; S23.2, Based on the calculated comprehensive health risk value The risk stratification of lighting indicators was completed, classifying blue light and glare indicators as high-risk health indicators, and illuminance indicators as routine comfort indicators.

[0010] As a further improvement to this technical solution, the adaptive fusion processing unit includes a risk reception and neighborhood adjustment module; the risk reception and neighborhood adjustment module is used to receive risk stratification results and perform threshold neighborhood sensitivity and sampling parameter adjustment; the data reception and parameter adjustment process of the risk reception and neighborhood adjustment module includes the following steps: S31.1 Receive the risk stratification results output by the integrated risk coupling and stratification module, and enable the heterogeneous data priority fusion logic; S31.2 Activate the health threshold neighborhood sensitivity adjustment mechanism to increase the sampling frequency and weight gain slope of blue light and glare high-risk indicators in the critical health state. .

[0011] As a further improvement to this technical solution, the adaptive fusion processing unit further includes a differentiated weight adjustment module; the differentiated weight adjustment module is used to set weight rules for different types of lighting indicators; the indicator weight adjustment process of the differentiated weight adjustment module includes the following steps: S32.1 Apply weighted gain adjustment to high-risk health indicators such as blue light and glare, and configure the weighted gain coefficient. ; S32.2 Apply weighted constraints to the conventional comfort index of illuminance and configure the weighted constraint coefficients. .

[0012] As a further improvement to this technical solution, the adaptive fusion processing unit further includes a differential normalization fusion module; the differential normalization fusion module is used to complete data normalization, multi-indicator fusion, and output feature data; the data normalization and fusion output process of the differential normalization fusion module includes the following steps: S33.1. Differential normalization processing is completed by preserving the boundary characteristics of the health threshold, thereby achieving the fusion of multiple types of indicator data; S33.2 Output the high-precision lighting environment feature data obtained by fusion. .

[0013] As a further improvement to this technical solution, the healthy lighting compliance decision-making unit includes an environmental characteristic compliance verification module, an operating condition information matching module, and a lighting instruction generation module, wherein: The environmental feature compliance verification module is used to receive high-precision lighting environment feature data. The built-in classroom healthy lighting threshold judgment model is used to complete the compliance verification of lighting data; The working condition information matching module is used to collect and match the real-time personnel distribution status in the classroom with the on-site environmental working conditions. The lighting instruction generation module is used to combine compliance verification results with real-time teaching environment conditions to generate healthy lighting control instructions that are adapted to the teaching environment.

[0014] As a further improvement to this technical solution, the intelligent lighting adaptive drive control unit includes a control command receiving and parsing module, a high-frequency PWM drive execution module, and a global lighting dynamic adjustment module, wherein: The control command receiving and parsing module is used to receive health lighting control commands and complete command recognition and adaptation parsing. The high-frequency PWM drive execution module outputs a high-frequency PWM stepless drive signal based on the parsed control instructions. The global lighting dynamic adjustment module relies on a high-frequency PWM drive signal to dynamically adjust the illuminance, color temperature, blue light ratio, and glare suppression of the classroom's global lighting terminal equipment.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention improves the assessment dimensions of lighting health risks by using multi-source sensor data time-series registration and coarse noise reduction preprocessing, constructing an instantaneous health risk assessment function, using a sliding time window to statistically analyze the time-cumulative risk factors of blue light and glare, and weighted coupling of instantaneous health risks and time-cumulative risks to obtain comprehensive health risks and complete the risk stratification of lighting indicators. It takes into account both the instantaneous state and the cumulative impact of long-term teaching scenarios, and enhances the comprehensiveness of health risk assessment. 2. This invention achieves dynamic adaptation of multi-indicator control weights and response strategies by using the following technical means: enabling heterogeneous data priority fusion logic based on risk stratification results, activating a health threshold neighborhood sensitivity adjustment mechanism, applying weight gain adjustment to high-risk indicators of blue light and glare, applying weight constraint adjustment to conventional comfort indicators of illuminance, and employing differentiated normalization processing that preserves the boundary characteristics of health thresholds. This optimizes the control response characteristics under different teaching conditions and improves the situation of rigid adjustment and control lag. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the overall system framework of the present invention; The meanings of the labels in the diagram are as follows: 1. Multi-source sensor synchronous acquisition unit; 11. Sensor acquisition module; 12. Data synchronization processing module; 13. Time sequence marker generation module; 2. Health Risk Analysis Unit; 21. Preprocessing and Instantaneous Risk Calculation Module; 22. Time-Cumulative Risk Calculation Module; 23. Comprehensive Risk Coupling and Hierarchy Module; 3. Adaptive fusion processing unit; 31. Risk reception and neighborhood adjustment module; 32. Differentiated weight adjustment module; 33. Differentiated normalization fusion module; 4. Healthy Lighting Compliance Decision-Making Unit; 41. Environmental Characteristic Compliance Verification Module; 42. Operating Condition Information Matching Module; 43. Lighting Instruction Generation Module; 5. Intelligent lighting adaptive drive control unit; 51. Control command receiving and parsing module; 52. High-frequency PWM drive execution module; 53. Global lighting dynamic adjustment module. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] like Figure 1 As shown, this embodiment provides a smart classroom health lighting system based on multi-sensor fusion, including: The multi-source sensor synchronous acquisition unit 1 collects various types of raw environmental data within the smart classroom, performs synchronous processing and timestamping, and generates a raw sensor dataset with time-series identifiers. Specifically, it includes a sensor acquisition module 11, a data synchronization processing module 12, and a time-series identifier generation module 13, wherein: The sensor acquisition module 11 collects various types of raw environmental data within the smart classroom. Specifically, the sensor acquisition module 11 integrates multiple types of sensors, including an illuminance sensor, a color temperature sensor, a spectral analysis sensor, and a glare monitoring camera module. The illuminance sensor is deployed on the classroom desktop and vertical work surfaces to collect illuminance data; the color temperature sensor collects the color temperature parameters of the ambient light source; the spectral analysis sensor focuses on collecting blue light radiation intensity data in the 415nm to 455nm wavelength range; the glare monitoring camera module obtains raw data related to the unified glare index by analyzing the ratio of high-brightness light source brightness to background brightness within the field of view. Each sensor operates independently, outputting raw data streams according to its respective sampling frequency.

[0019] The data synchronization processing module 12 performs synchronization processing on various types of raw environmental data collected by the sensor acquisition module 11. Specifically, due to differences in hardware response delays, internal data output rates, and communication protocols among different sensors, the data from each channel is not synchronized in time. The data synchronization processing module 12 achieves data synchronization through two methods: hardware synchronization triggering and software interpolation alignment. In the hardware synchronous triggering mode, the data synchronization processing module 12 uses a shared synchronous clock source to perform hardware triggering acquisition of the spectral analysis sensor and the glare monitoring camera module, achieving microsecond-level time alignment from the source. In the software interpolation alignment mode, the data synchronization processing module 12 has a built-in high-speed cache area to cache the data of each channel in the most recent time window. Taking the time axis of the illuminance sensor with the highest data update frequency as the reference, it performs linear interpolation or nearest neighbor matching on the spectral data and glare data with lower update frequency, thereby reconstructing the discrete data stream of all channels into a multi-dimensional data vector with synchronous consistency at a unified time point.

[0020] The timing stamp generation module 13 timestamps the data after synchronization processing by the data synchronization processing module 12, generating a raw sensor dataset with timing identifiers. Specifically, the timing stamp generation module 13 is equipped with a local high-stability crystal oscillator clock that is periodically calibrated with the network time server, achieving millisecond-level timestamp accuracy. Upon receiving a synchronization data frame, the timing stamp generation module 13 latches the current precise time value, encapsulates this time value in Unix timestamp or ISO 8601 format into the data frame header, and appends a monotonically increasing frame sequence number. After this processing, each frame of synchronization data carries a precise timing identifier, forming a raw sensor dataset with timing identifiers.

[0021] This original sensor dataset provides a reliable time reference for calculating the time-cumulative risk of indicators such as blue light and glare in the subsequent health risk analysis unit 2, eliminating the time deviation introduced by the asynchronous sampling of sensors and system processing.

[0022] Health risk analysis unit 2 receives the original sensor dataset and sequentially performs temporal registration and coarse noise reduction processing. Based on preset health thresholds for various sensor indicators, it constructs an instantaneous health risk assessment function, calculates the time-cumulative risk factors corresponding to blue light and glare indicators, couples the instantaneous health risk with the time-cumulative risk factors to obtain the comprehensive health risk, and completes the risk stratification of lighting indicators based on the comprehensive health risk. Specifically, it includes a preprocessing and instantaneous risk calculation module 21, a time-cumulative risk calculation module 22, and a comprehensive risk coupling and stratification module 23, wherein: The preprocessing and instantaneous risk calculation module 21 is used to clean and preprocess the original sensor dataset and quantify the instantaneous health risks of various lighting indicators. The data preprocessing and instantaneous risk calculation process of the preprocessing and instantaneous risk calculation module 21 includes the following steps: S21.1 Receive the raw sensor dataset, perform time registration processing and coarse noise reduction processing in sequence, unify the time reference of all sensor data, and remove abnormal data in the dataset that exceed the physical range. In this step, the specific process of time-series registration is as follows: The preprocessing and instantaneous risk calculation module 21 checks the frame number and timestamp of each frame of data in the original sensor dataset one by one to identify whether there are any anomalies such as timestamp jumps, out-of-order or duplications caused by network transmission delays or data caching. When uneven timestamp intervals are detected, the preprocessing and instantaneous risk calculation module 21 performs linear interpolation resampling on the time axis based on a preset standard sampling interval, thereby unifying the time reference of all sensor data and ensuring that the data from different sensor channels are strictly aligned in the time dimension.

[0023] The coarse noise reduction process employs a physical range constraint method. The preprocessing and instantaneous risk calculation module 21 internally stores the effective physical ranges of various types of sensors, including: 0 lx to 20000 lx for illuminance sensors, 1000 K to 15000 K for color temperature sensors, and 0 μW / cm² for blue light radiation intensity in the 415 nm to 455 nm wavelength band for spectral analysis sensors. 2 Up to 1000 μW / cm 2 The effective range of the unified glare index for the glare monitoring camera module is 10 to 30. The preprocessing and instantaneous risk calculation module 21 iterates through each sampled value from each sensor channel in the dataset, comparing the sampled value with the effective physical range of the corresponding sensor. If the sampled value falls within the effective range, it is retained; if the sampled value exceeds the effective range, it is determined as abnormal data and discarded, and the position of that channel at that moment is marked as null.

[0024] S21.2 Construct an instantaneous health risk assessment function based on preset health thresholds for various sensor indicators, and calculate and output the instantaneous health risk values ​​corresponding to various sensor indicators. The subscript Indicates the type of sensing index, in this embodiment .

[0025] In this step, considering the different mechanisms by which various lighting indicators affect human health, this embodiment employs a composite evaluation model: the illuminance indicator, due to the existence of an optimal comfort range, uses a piecewise linear function; the color temperature indicator deviates symmetrically around its optimal center value, using an S-shaped function; blue light and glare are unilateral risk indicators, employing exponential and logistic functions respectively. This differentiated design avoids abrupt risk changes near the threshold caused by traditional uniform function forms, ensuring smooth control commands. The specific calculation formulas for each indicator are given below.

[0026] Instantaneous health risk value of illuminance index: The further the illuminance deviates from the optimal comfort range, the higher the health risk. The optimal lower limit threshold for illuminance is defined as... The optimal upper limit threshold is The completely unacceptable lower threshold is The completely unacceptable upper limit threshold is Instantaneous health risk values ​​are calculated using a piecewise linear function: ; in: This represents the instantaneous health risk value of the illuminance index. It is dimensionless and ranges from [0,1]. The larger the value, the higher the instantaneous health risk. This represents the pre-processed illuminance measurement value at the current moment, in lux (lx). The optimal lower limit health threshold for illuminance is represented by 300 lx in this embodiment; The optimal upper limit health threshold for illuminance is represented; in this embodiment, it is set to 500 lx. This represents the critical lower limit threshold of illuminance. Values ​​below this threshold are considered completely unacceptable. In this embodiment, it is set to 100 lx. This represents the critical upper limit threshold of illuminance. Values ​​exceeding this threshold are considered completely unacceptable. In this embodiment, it is set to 2000 lx.

[0027] Instantaneous health risk value of color temperature index: Both excessively high and low color temperatures can cause visual fatigue and disrupt physiological rhythms. The optimal center threshold for color temperature is defined as... The acceptable deviation threshold is The limit deviation threshold is A symmetric S-shaped risk function is used: ; in: This represents the instantaneous health risk value of the color temperature index, is dimensionless, and ranges from (0,1). When the risk value approaches 0, The risk value is 0.5 when The risk value approaches 1. This indicates the color temperature measurement value at the current moment, in Kelvin (K). The optimal health threshold for color temperature is 4500 K in this embodiment; This represents the acceptable deviation threshold for color temperature, which is set to 500 K in this embodiment. The threshold value representing the extreme deviation of color temperature is taken as 1500 K in this embodiment; This represents the sensitivity adjustment coefficient of the color temperature risk function, used to control the steepness of the S-curve in the transition zone. In this embodiment, it is taken as... .

[0028] Instantaneous health risk values ​​of blue light levels: The higher the intensity of blue light radiation (focusing on the 415nm to 455nm band), the greater the potential damage to the retina and circadian rhythms; this is a unilateral risk indicator. The safe threshold for blue light is defined as follows: The danger threshold is An exponential risk function is used, resulting in a gradual increase in risk near the safety threshold, followed by a rapid increase beyond that threshold. ; in: This represents the instantaneous health risk value of the blue light index, which is dimensionless and ranges from [0,1]. This represents the measured value of blue light radiation intensity at the current moment, expressed in microwatts per square centimeter (μW / cm²). 2 ); This represents the safety and health threshold for blue light; below this value, the instantaneous risk is considered zero. In this embodiment, 100 μW / cm² is used. 2 ; This represents the dangerous health threshold of blue light; reaching or exceeding this value is considered to pose the greatest instantaneous risk. In this embodiment, 500 μW / cm² is used. 2 ; This represents the growth rate coefficient of the blue light risk function, and is taken as a positive value. In this embodiment, it is taken as... , dimensionless.

[0029] when near When the risk value rapidly approaches 1, in order to simplify the embedded system control logic, in this embodiment... Take directly 1.

[0030] Instantaneous health risk value of glare index: A higher Uniform Glare Index (UGR) indicates greater visual discomfort and is also a one-sided risk indicator. The acceptable threshold for glare is defined as... The unacceptable threshold is An improved Logistic function is used, which exhibits high sensitivity near the threshold. ; in: This represents the instantaneous health risk value of the glare index, which is dimensionless and ranges from (0,1). When the risk value is less than 0.5, The risk value is 0.5 when The risk value is approximately 0.95. Represents the current uniform glare index measurement, dimensionless; This represents the acceptable health threshold for glare, which is set to 16 in this embodiment; The unacceptable health threshold for glare is represented by 22 in this embodiment; The gain coefficient representing the glare risk function is based on... The risk value is determined by reverse calculation based on the condition that it equals 0.95. In this embodiment... Take 0.49, which is dimensionless.

[0031] At this point, the preprocessing and instantaneous risk calculation module 21 calculates the instantaneous health risk values ​​of illuminance, color temperature, blue light, and glare indicators sequentially using the four formulas given above. Collectively referred to as And output it to subsequent modules.

[0032] The core improvements to the aforementioned evaluation function are: a heterogeneous function form was designed to address the physiological and physical characteristics of different indicators, rather than using a uniform mathematical expression; a continuously differentiable transition region was introduced at the health threshold boundary, avoiding abrupt changes in control commands caused by traditional step functions; and an adjustable sensitivity parameter enabled engineering adaptation to differentiated response speeds for different risk indicators. All preset health thresholds can be reconfigured online according to actual teaching scenarios and the latest lighting health standards, ensuring the system's scalability and regulatory compatibility.

[0033] The time-cumulative risk calculation module 22 is used to quantify the time-series cumulative risk of high-risk lighting indicators based on instantaneous health risks. Since blue light radiation has a dose-cumulative effect on the visual system, and prolonged glare exposure can exacerbate visual fatigue, assessing only instantaneous risk is insufficient to comprehensively characterize the health impact. The time-cumulative risk calculation module 22 incorporates the time dimension into risk assessment through a sliding time window statistical model, enabling a reasonable determination of "short-term tolerance for exceeding limits and long-term cumulative warning" for high-risk indicators. The calculation process of the time-cumulative risk factor in the time-cumulative risk calculation module 22 includes the following steps: S22.1 Receive the instantaneous health risk value output by the preprocessing and instantaneous risk calculation module 21. From this, values ​​corresponding to two categories of indicators, blue light and glare, are selected. In this step, the time-cumulative risk calculation module 22 first receives all instantaneous health risk values ​​output by the preprocessing and instantaneous risk calculation modules 21. ,in Then, based on the pre-set high-risk indicator labels in health risk analysis unit 2, the instantaneous health risk value corresponding to the blue light indicator is selected. Instantaneous health risk value corresponding to glare index This serves as the input for subsequent time-cumulative calculations. Due to their different physical and physiological characteristics, illuminance and color temperature are not included in the time-cumulative risk calculation in this embodiment to avoid overly conservative control.

[0034] S22.2 Activate the preset sliding time window statistical mode, combining the time series data within the window with the instantaneous health risk value. Calculate and output the time-cumulative risk factor of the corresponding indicator. .

[0035] In this step, for the two high-risk indicators identified—blue light and glare—the time-cumulative risk calculation module 22 employs independent sliding time window statistical modes. The length of each window is W (in seconds), and the sliding step size is... (That is, the time interval between two consecutive cumulative risk calculations is equal to the system's main control cycle). The time-cumulative risk calculation module 22 internally maintains a first-in-first-out circular buffer to store the instantaneous health risk value sequence at each moment within the window.

[0036] On the one hand, regarding blue light indicators, there is a cumulative risk factor over time. The calculation uses an exponentially weighted moving average model, which assigns higher weight to recent instantaneous risks, consistent with the "recency effect" in the cumulative physiological response of the human body. The calculation formula is as follows: ; in: Indicates the current time The time-cumulative risk factor of the blue light index is dimensionless and its value ranges from [0,1]. This indicates the current calculation time, in seconds. This indicates the number of instantaneous risk sampling points included within the sliding time window. ,symbol Indicates rounding down; This represents the main control cycle of the system, which is the interval between two consecutive cumulative risk calculations. In this embodiment, it is taken as 1 second. This indicates the length of the sliding time window, which is 300s (i.e., 5 minutes) in this embodiment. For sampling point index, Indicates the current moment. This indicates the previous control cycle, and so on. Indicates the past number The instantaneous health risk value of blue light at each control cycle moment; This represents the forgetting factor, with a value range of (0,1]. The smaller the value, the faster the contribution of long-term historical data to the current accumulated risk decays. This embodiment emphasizes the recency effect by taking... .

[0037] On the other hand, for glare indicators, the time-cumulative risk factor The calculation employs a threshold-based continuous integral model, which focuses on the duration after the instantaneous glare risk exceeds a certain tolerance threshold. This model is applicable to the increased discomfort caused by the accumulation of the Unified Glare Index (UGR) over time. The calculation formula is: ; in: Indicates the current time The time-cumulative risk factor of the glare index is dimensionless and its value ranges from [0,1]. This represents the integral variable, indicating any moment within the window, in seconds. This indicates an indicator function that takes the value 1 when the condition inside the parentheses is true, and 0 otherwise. The tolerance threshold representing the instantaneous risk of glare is taken in this embodiment. =0.5, meaning that moments with instantaneous risk exceeding 0.5 are included in the cumulative exposure time; As before, use 300s.

[0038] In practical discrete control systems, the above integral is approximated as a summation: .

[0039] Based on the above calculations, the time-cumulative risk calculation module 22 outputs the time-cumulative risk factor of the blue light index. Time-cumulative risk factors for glare indicators For the sake of consistency, In this embodiment, the corresponding time-cumulative risk factor for the blue light index is... Time-cumulative risk factors for glare indicators subscript It implies the accumulation of time and can also distinguish different types of indicators.

[0040] The core improvement of this module lies in the design of exponentially weighted moving average models and threshold continuous integral models for the different physiological pathways of blue light and glare, instead of using a single cumulative algorithm; and the length of the sliding time window. With sliding step size Independently configurable, balancing computational load and response sensitivity; forgetting factor in the blue light accumulation model. This enables the system to respond more quickly to recent changes in risk without ignoring the contribution of longer historical exposures; the tolerance threshold in the glare accumulation model. This avoids the excessive accumulation of minor, transient fluctuations, reducing the probability of control malfunctions. All parameters ( , , , All of these can be adjusted online according to the actual usage scenarios and health standards of the classroom, ensuring the flexibility and adaptability of the system.

[0041] The integrated risk coupling and stratification module 23 is used to fuse instantaneous risk and cumulative risk, and to classify lighting health risk levels. Since blue light and glare have a cumulative damage effect over time, while illuminance and color temperature mainly affect instantaneous visual comfort, this module adopts a differentiated weighted fusion strategy: for indicators with cumulative risk (blue light, glare), the integrated risk is jointly determined by instantaneous risk and cumulative risk; for indicators without cumulative risk (illuminance, color temperature), the integrated risk is directly taken as the instantaneous risk value. The risk coupling and indicator stratification process of the integrated risk coupling and stratification module 23 includes the following steps: S23.1 Receive instantaneous health risk value With time-cumulative risk factors The comprehensive health risk value of each lighting indicator is obtained by coupling through a weighted fusion method. ; In this step, for illuminance and color temperature, since they are not included in the calculation of cumulative risk over time, their comprehensive health risk value is directly equal to the instantaneous health risk value. For blue light and glare, a weighted summation method is used to couple instantaneous risk and cumulative risk, and the calculation formula is as follows: ; ; ; ; in: This represents the comprehensive health risk value of blue light indicators. It is dimensionless and ranges from [0,1]. The comprehensive health risk value of the glare index is dimensionless and ranges from [0,1]. This represents the comprehensive health risk value of the illuminance index, which is dimensionless and ranges from [0,1]. The color temperature index represents the overall health risk value, which is dimensionless and ranges from [0,1]. The fusion weight coefficient represents instantaneous risk; it is dimensionless and ranges from [0,1]. The fusion weight coefficient represents the cumulative risk; it is dimensionless, takes values ​​in the range [0,1], and satisfies the following conditions: .

[0042] In this embodiment, considering the long-term health effects of blue light and glare, the following settings are provided. This means that the cumulative risk weight is slightly higher than the instantaneous risk weight to strengthen the constraint on long-term exposure. The aforementioned weighting coefficients can be dynamically adjusted according to the continuous usage duration of the teaching scenario; for example, they can be automatically increased when the continuous class time exceeds 2 hours. Increase the proportion of decisions based on accumulated risks.

[0043] S23.2, Based on the calculated comprehensive health risk value The risk stratification of lighting indicators was completed, classifying blue light and glare indicators as high-risk health indicators, and illuminance indicators as routine comfort indicators.

[0044] In this step, the integrated risk coupling and stratification module 23 calculates the integrated health risk value of each indicator based on step S23.1. Based on expertise in the field of lighting health, the four lighting indicators are divided into two risk levels: High-risk health indicator layer: This layer includes blue light and glare indicators. The criteria for judgment are as follows: these two types of indicators not only have instantaneous health effects, but also have a clear cumulative damage effect over time. Once their combined health risk value exceeds the preset alarm threshold, it may cause retinal photochemical damage or severe visual fatigue. Therefore, in the subsequent adaptive fusion processing unit 3, differentiated priority fusion logic will be used for the blue light and glare indicators, including enhanced processing such as increasing the sampling frequency and applying weight gain adjustment.

[0045] Standard comfort index level: This layer includes illuminance and color temperature indices. The criteria for selection are as follows: these two indices primarily affect visual comfort and task performance. Although deviations from the optimal range can cause discomfort, they generally do not possess irreversible cumulative damage characteristics, and the human body has a good ability to adapt to short-term fluctuations in ambient illuminance and color temperature. Therefore, in the subsequent adaptive fusion processing unit 3, the illuminance and color temperature indices undergo conventional weight constraint adjustments, without the need to activate special enhancement mechanisms.

[0046] The risk stratification results from the integrated risk coupling and stratification module 23 are output to the adaptive fusion processing unit 3 in the form of identifiers. Specifically, the system assigns a risk level label to each indicator. ,in: ; ; Meanwhile, the integrated risk coupling and hierarchical module 23 also outputs the comprehensive health risk value of each indicator. (Complete numerical values) and an overall risk indicator to trigger corresponding control strategies in subsequent units.

[0047] The core improvement of the integrated risk coupling and hierarchical module 23 lies in: adopting differentiated fusion rules based on whether different indicators have cumulative damage characteristics (indicators with cumulative risk are weighted and fused, while indicators without cumulative risk directly inherit instantaneous risk), thus avoiding redundant calculations or risk underestimation caused by unified fusion; fusion weight coefficients It can be dynamically adjusted according to the continuous usage time of the classroom, realizing the time adaptability of risk assessment; the risk stratification results do not rely on a single threshold judgment, but combine the physiological mechanism of the indicator (whether there is a cumulative effect), so that the stratification logic has a clear medical and engineering basis; the output includes both discrete risk level labels and retains continuous comprehensive risk values, providing rich information for the fine adjustment of the subsequent adaptive fusion processing unit 3.

[0048] After the above steps, the integrated risk coupling and hierarchical module 23 completes the full mapping from multi-indicator risk values ​​to risk hierarchies, and efficiently and unambiguously transmits the results to the adaptive fusion processing unit 3.

[0049] The adaptive fusion processing unit 3 receives the risk stratification results, carries heterogeneous data priority fusion logic, and enables a health threshold neighborhood sensitivity adjustment mechanism. Under critical health conditions, it increases the sampling frequency and weight gain slope of high-risk blue light and glare indicators, applies weight gain adjustment to high-risk blue light and glare health indicators, and applies weight constraint adjustment to conventional comfort indicators of illuminance. Data fusion is completed through differentiated normalization processing that preserves the boundary characteristics of health thresholds, outputting high-precision lighting environment characteristic data. Specifically, it includes a risk reception and neighborhood adjustment module 31, a differentiated weight adjustment module 32, and a differentiated normalization fusion module 33, wherein: The risk reception and neighborhood adjustment module 31 is used to receive the risk stratification results and perform threshold neighborhood sensitivity and sampling parameter adjustment; the data reception and parameter adjustment process of the risk reception and neighborhood adjustment module 31 includes the following steps: S31.1 Receive the risk stratification results output by the integrated risk coupling and stratification module 23, and enable the heterogeneous data priority fusion logic; In this step, the risk reception and neighborhood adjustment module 31 first receives output information from the integrated risk coupling and hierarchical module 23, which includes: Comprehensive health risk value of various lighting indicators ; Risk level labels for each lighting indicator The labels for blue light and glare are: The labels for illuminance and color temperature are as follows: .

[0050] Based on the aforementioned tags, the risk reception and neighborhood adjustment module 31 enables heterogeneous data priority fusion logic. The core idea of ​​this logic is that indicators with different risk levels should enjoy different processing priorities and resource allocation weights in subsequent fusion processing. Specifically, the module establishes two priority queues: High-priority queue: Stores blue light and glare index data, and enjoys priority processing rights in subsequent sampling, transmission and fusion processes; Regular priority queue: Stores illuminance and color temperature index data and processes them according to the regular workflow.

[0051] After enabling the heterogeneous data priority fusion logic, the risk reception and neighborhood adjustment module 31 will enter the dynamic adjustment preparation state in step S31.2.

[0052] S31.2 Activate the health threshold neighborhood sensitivity adjustment mechanism to increase the sampling frequency and weight gain slope of blue light and glare high-risk indicators in the critical health state. .

[0053] In this step, the risk reception and neighborhood adjustment module 31 continuously monitors the comprehensive health risk value of blue light and glare indicators. and And compare it with a preset critical health state trigger threshold. Compare them. When any high-risk indicator meets... When the system determines that it is currently in a "critical health state" for the indicator, it immediately activates the health threshold neighborhood sensitivity adjustment mechanism.

[0054] In this embodiment, the threshold for triggering a critical health state is set to... The value is set based on the following: when the comprehensive health risk value exceeds 0.7, it indicates that the lighting environment has approached or reached the upper limit of the health standard, and it is necessary to increase the monitoring sensitivity to avoid the risk getting out of control.

[0055] In a critically healthy state, the risk reception and neighborhood adjustment module 31 performs the following two adjustment operations: Increase the sampling frequency of high-risk indicators: For blue light or glare indicators that have entered a critical health state, the module sends a sampling frequency adjustment command to the corresponding sensor (spectral analysis sensor or glare monitoring camera module). The original conventional sampling frequency is... (In this embodiment, 1 Hz is used, i.e., 1 sampling per second). Under critical health conditions, the sampling frequency is increased to... The calculation formula is: ; in: The sampling frequency represents the critical health state, and the unit is Hertz (Hz). This indicates the maximum safe sampling frequency supported by the sensor hardware. In this embodiment, the spectral analysis sensor uses 10Hz, and the glare monitoring camera module uses 5Hz. This represents the sampling frequency boost factor, which is dimensionless. In this embodiment, it is taken as... This means that the sampling frequency is increased to 5 times the normal value, but does not exceed the hardware limit; This represents the standard sampling frequency, which is 1 Hz in this embodiment.

[0056] When the overall health risk value of this indicator falls back to After the following condition persists for more than 30 seconds (i.e., confirming that the critical state has been lifted), the module will automatically restore the sampling frequency to the normal value. .

[0057] Increase the slope of the weighted gain : Weighted gain slope It is used in the subsequent differential weight adjustment module 32 to calculate the weight gain coefficients for blue light and glare indicators. The key parameter is the weight gain slope, which takes the default value under normal conditions. In a critical health state, the module increases this slope to This is to enhance the sensitivity of the weight response of high-risk indicators in data fusion.

[0058] In this embodiment, The definitions and adjustment rules are as follows: ; in: This represents the slope of the weight gain, which is dimensionless and used to calculate the weight gain coefficient of high-risk indicators in the differential weight adjustment module 32. ; This represents the default weight gain slope under normal conditions; in this embodiment, it is taken as... ; This represents the slope of the target weight gain under the critical health state; in this embodiment, it is taken as... ; This represents the maximum allowable value of the weighted gain slope; in this embodiment, it is taken as... ; This represents the width of the neighborhood transition band, used to avoid frequent fluctuations in risk values ​​near the threshold that could cause parameter jumps. In this embodiment, it is taken as... ; This represents the overall health risk value of blue light or glare indicators. It is dimensionless and ranges from [0,1].

[0059] The above rule achieves a smooth increase in the slope of the weighted gain as the overall health risk value increases: when the risk value is lower than... When the risk value is 0.65, maintain the default slope of 1.0; when the risk value enters... When the risk value is within the range of [1.0, 3.0], the slope increases linearly; when the risk value reaches or exceeds... When the slope is set to the target value of 3.0, the design ensures both low sensitivity when the risk value is far from the threshold (avoiding over-adjustment) and high sensitivity when the risk value is close to the threshold (achieving fine control).

[0060] After the adjustment in step S31.2, the risk reception and neighborhood adjustment module 31 will update the parameters—including the adjusted sampling frequency. (or ), weighted gain slope --The risk stratification results are also transmitted to the differential weight adjustment module 32. At the same time, the risk receiving and neighborhood adjustment module 31 maintains a status flag to record whether each high-risk indicator is in a critical health state, so as to restore normal parameters in a timely manner after leaving the critical state.

[0061] The core improvements of this module are: the introduction of an explicit determination mechanism for "critical health state," which differs from the traditional system's approach of relying solely on a single threshold for on / off control; continuous adjustment of the sampling frequency and weighted gain slope based on risk value, rather than simple binary switching, thus avoiding control chattering; and the introduction of a neighborhood transition band. Setting a smooth transition range near the threshold improves the system's robustness; enabling heterogeneous data priority fusion logic ensures that high-priority indicators are protected during resource contention (such as communication bandwidth and computing resources), guaranteeing low-latency processing of critical data. All adjustment parameters ( All of them can be configured online according to classroom lighting health standards and actual application needs, and have good engineering adaptability.

[0062] The differentiated weight adjustment module 32 is used to set weight rules for different types of lighting indicators; the indicator weight adjustment process of the differentiated weight adjustment module 32 includes the following steps: S32.1 Apply weighted gain adjustment to high-risk health indicators such as blue light and glare, and configure the weighted gain coefficient. ; In this step, the differential weighting adjustment module 32 first identifies the risk level label as... The metrics are blue light and glare. For these two metrics, the module employs a weighted gain adjustment strategy, calculating a weighted gain coefficient greater than or equal to 1 for each. This coefficient will be used in the subsequent data weighting fusion process in the differential normalization fusion module 33.

[0063] Weighted gain coefficient The calculation is based on the weighted gain slope provided by the risk reception and neighborhood adjustment module 31. And combined with the comprehensive health risk value of this indicator at the current moment. The calculation formula is as follows: ; in: Indicates the first The weighting gain coefficients for high-risk health indicators are dimensionless and range from [1, 2]. When the overall health risk value is 0, the gain coefficient is 1 (no gain); when the overall health risk value is 1, the gain coefficient is... (Maximum gain); The slope of the weight gain is dynamically adjusted by the risk receiving and neighborhood adjustment module 31 in step S31.2. It is 1.0 under normal conditions and can be increased to 3.0 under critical health conditions. This represents a non-linear adjustment index, dimensionless, used to control the shape of gain growth with risk value. When... At that time, gains in low-risk areas increased more rapidly; when At that time, the gain increases faster in high-risk areas. This embodiment aims to enhance the sensitivity of high-risk areas by taking... .

[0064] In this embodiment, the weighted gain coefficients for blue light and glare indicators are calculated independently, denoted as... and For example, at a certain moment ,and (In a borderline state of health), then: ; This calculation means that the weight of the blue light index in subsequent fusion will be amplified to approximately 2.92 times the original weight.

[0065] S32.2 Apply weighted constraints to the conventional comfort index of illuminance and configure the weighted constraint coefficients. .

[0066] In this step, the differential weight adjustment module 32 simultaneously identifies the indicators with the risk level label ROUTINE_COMFORT, namely illuminance and color temperature. For these two types of indicators, the differential weight adjustment module 32 adopts a weight constraint adjustment strategy, calculating a weight constraint coefficient with a value range of [0,1] for each. The introduction of this coefficient aims to reasonably limit the weight of conventional comfort indicators in data fusion, and to prevent them from excessively dominating the fusion results in non-critical states.

[0067] Weight constraint coefficient The calculation is based on the comprehensive health risk value of conventional comfort indicators. The calculation employs a monotonically decreasing function: when the overall health risk value is low, the constraint coefficient approaches 1 (i.e., virtually no constraint is applied); when the overall health risk value is high, the constraint coefficient is appropriately reduced to decrease the fusion weight of this indicator and prevent it from masking the impact of high-risk indicators. The calculation formula is as follows: ; in: Indicates the first The weight constraint coefficients for conventional comfort indicators are dimensionless and range from [ ]. [1] In this embodiment, it is ensured ; Indicates the first The comprehensive health risk value of the conventional comfort index is dimensionless and ranges from [0,1]. Its definition has been given in step S23.1. This represents the maximum constraint amplitude, which is dimensionless and ranges from (0,1]. In this embodiment, it is taken as... That is, the constraint coefficient can be reduced to a minimum of 0.7; This represents the shape exponent of the constraint curve; it is dimensionless and takes positive values. When At that time, the constraint effect of low-risk areas is relatively obvious; when At that time, the constraint effect of high-risk areas is more significant. In this embodiment, to maintain the stability of conventional comfort indicators within the vast majority of normal ranges, [the following is taken:] (The quadratic relationship) ensures that the constraint coefficient decreases significantly only when the risk value is high.

[0068] In this embodiment, the illuminance and color temperature indices are calculated independently using their respective weighting constraint coefficients, denoted as... and For example, at a certain moment (If the illuminance deviates significantly from the optimal range), then: ; This calculation means that the weight of the illuminance index in subsequent fusion will be multiplied by approximately 0.757, which is equivalent to applying a constraint attenuation of approximately 24.3%. When the illuminance risk value is low (e.g., ... )hour, Almost unrestricted.

[0069] Finally, after completing the above calculations, the differential weight adjustment module 32 outputs the following parameters to the differential normalization fusion module 33: Weighting gain coefficient of high-risk health indicators , ; Weighting constraint coefficients for conventional comfort indicators , ; The raw measurement data of each indicator (preprocessed illuminance, color temperature, blue light radiation intensity, and uniform glare index) and their corresponding comprehensive health risk values. .

[0070] The core improvement of this module lies in designing two modes—gain adjustment and constraint adjustment—separately for high-risk and regular indicators, instead of using a uniform weighting strategy for all indicators; the weight gain coefficient... It not only relies on risk level, but also incorporates dynamically adjusted weighted gain slope. With nonlinear exponent This achieves a continuous mapping from "critical state perception" to "fusion weight response"; weight constraint coefficients A monotonically decreasing function is used to ensure that conventional indicators appropriately reduce their weight under high-risk conditions, thus avoiding diluting key health information; all adjustment parameters ( , , Each component is independently configurable, facilitating customized optimization for different teaching scenarios (such as experimental classes, reading classes, and self-study classes). Through the aforementioned differentiated weight adjustments, the system maximizes the reference value of environmental comfort data while ensuring basic lighting health standards.

[0071] The differential normalization fusion module 33 is used to complete data normalization, multi-indicator fusion, and output feature data; the data normalization and fusion output process of the differential normalization fusion module 33 includes the following steps: S33.1. Differential normalization processing is completed by preserving the boundary characteristics of the health threshold, thereby achieving the fusion of multiple types of indicator data; In this step, the differential normalization fusion module 33 employs different normalization mapping functions for the four indicators: illuminance, color temperature, blue light, and glare. The common design principle for all mapping functions is to define the health threshold boundaries of each indicator (i.e., the various thresholds defined in step S21.2, such as the optimal range boundary for illuminance). , Blue light safety threshold (etc.) is mapped to a normalization value of 0.5, and the absolute value of the derivative of the normalization function is maximized at the threshold boundary, thereby ensuring that small changes in physical quantities near the health boundary can cause significant changes in the normalization value.

[0072] Differential normalization of illuminance index: The health-sensitive region of illuminance indices is concentrated near the boundary of the optimal interval. Define the normalized illuminance characteristic components. The calculation formula is: ; in: This represents the normalized health characteristic component of the illuminance index, is dimensionless, and ranges from (0,1). When the illuminance falls within the optimal interval... hour It approaches the maximum value of 1, and approaches 0 when the illuminance deviates significantly; This represents the current illuminance measurement, in lux (lx). The optimal lower limit health threshold for illuminance is represented by 300 lx in this embodiment; The optimal upper limit health threshold for illuminance is represented; in this embodiment, it is set to 500 lx. This represents the steepness coefficient of the illuminance normalization function, in lx. -1 This is used to control the gradient magnitude at the threshold boundary. In this embodiment, we take... lx -1 This makes in and Place It decreases to around 0.5, and the absolute value of the derivative is largest at the boundary.

[0073] The physical meaning of this formula is that the product of two Sigmoid functions constructs a "soft interval" window function, which takes a value close to 1 within the optimal interval and decays rapidly to 0 outside the interval, with the region of fastest decay located exactly on the boundary of the health threshold.

[0074] Differential normalization of color temperature index: The health-sensitive area of ​​the color temperature index is concentrated in the optimal center threshold. The closer to the target color temperature, the lower the normalized value. Define the normalized color temperature feature components. The calculation formula is: ; in: This represents the normalized health characteristic component of the color temperature index, which is dimensionless and ranges from (0,1]. The maximum value of 1 is taken when the value deviates from the target value, and the value is decayed according to the Gaussian function when it deviates from the target value. This indicates the color temperature measurement value at the current moment, in Kelvin (K). The optimal health threshold for color temperature is 4500 K in this embodiment; The bandwidth parameter of the Gaussian kernel, measured in Kelvin (K), controls the decay rate. This embodiment is based on an acceptable deviation threshold. The setting makes it possible for when hour ,Pick .

[0075] The Gaussian function is The absolute value of the derivative is 0 at the threshold, but the second derivative is the largest, actually at the threshold boundary (e.g. It has a relatively fast decay rate and retains boundary sensitivity.

[0076] Differential normalization of blue light index: Blue light is a one-sided risk indicator, and the health-sensitive area is concentrated around the safety threshold. Nearby. Define the normalized blue light characteristic components. The calculation formula is: ; in: This represents the normalized health characteristic component of the blue light index, which is dimensionless and ranges from (0,1). hour Approaching 0 (healthy), when hour Approaching 1 (dangerous); This represents the measured value of blue light radiation intensity at the current moment, expressed in microwatts per square centimeter (μW / cm²). 2 ); This represents the safe and healthy threshold for blue light; in this embodiment, it is taken as 100 μW / cm. 2 ; This represents the steepness coefficient of the blue light normalization function, in cm. 2 / μW, in this embodiment, is taken as This makes in Place And the derivative is maximum at that point.

[0077] This formula essentially inverts the Logistic function, mapping it to 0.5 at the safety threshold and achieving the highest sensitivity at that point.

[0078] Differential normalization of glare index: Glare index is also a one-sided risk indicator, with the health-sensitive area concentrated within the acceptable threshold. Nearby. Define normalized glare characteristic components. The calculation formula is: ; in: The normalized health characteristic component of the glare index is dimensionless and ranges from (0,1). hour Less than 0.5 (relatively healthy), when hour Approaching 1 (dangerous); Represents the current uniform glare index measurement, dimensionless; This represents the acceptable health threshold for glare, which is set to 16 in this embodiment; The steepness coefficient of the glare normalization function is dimensionless. In this embodiment, it is based on an unacceptable threshold. The setting makes it possible to... hour 0.95, derived backwards Take 0.49.

[0079] This formula uses the standard Logistic function, in Place It has a value of 0.5 and the derivative is the largest, thus preserving the sensitivity of the threshold boundary.

[0080] After completing the above differential normalization, four normalized feature components are obtained. They have all been unified to the [0,1] dimension and all have the largest local gradient at their respective health threshold boundaries.

[0081] S33.2 Output the high-precision lighting environment feature data obtained by fusion. .

[0082] Specifically, the differential normalization fusion module 33 will use the weight gain coefficients calculated in steps S32.1 and S32.2 to perform the fusion. (Regarding blue light and glare) and weighting constraint coefficients (For illuminance and color temperature) the corresponding normalized feature components are applied, and then a weighted sum is performed to obtain high-precision lighting environment feature data. The calculation formula is: ; It should be noted that for high-risk indicators such as blue light and glare, they enjoy both gain coefficients and other benefits. It is also subject to conventional constraint coefficients (but here, to maintain the symmetry of weight adjustment, a certain value can be set). This means that no constraint attenuation is imposed on high-risk indicators; if it is desired to impose constraints on high-risk indicators in certain scenarios, this can be achieved through configuration. In this embodiment, it is agreed that... Therefore, the above formula simplifies to: ; in: This represents the final output of high-precision lighting environment characteristic data. It is dimensionless and ranges from [0,1]. The closer the value is to 1, the more the lighting environment deviates from a healthy state (dangerous). The closer the value is to 0, the healthier the lighting environment is. , These represent the weight constraint correction coefficients for the high-risk indicators of blue light and glare, respectively. They are dimensionless and range from [0,1]. In this embodiment, when all indicators are in an ideal healthy state, and gain coefficient Approximately 1, constraint coefficient Approximately 1, at this time This is not 0, because the optimal states of illuminance and color temperature are mapped to 1, while the optimal states of blue light and glare are mapped to 0. After fusion, it needs to be reinterpreted according to the actual weights. For easier intuitive understanding, the system can output a separate health level rating, but... As continuous feature data, it is sufficient for subsequent compliance decision-making units to use.

[0083] The module outputs Together with each normalized feature component and weighting coefficient, it is transmitted to the Healthy Lighting Compliance Decision Unit 4 as high-precision lighting environment feature data.

[0084] The core improvement of this module lies in: designing a differentiated normalization function that "preserves the boundary characteristics of the health threshold," ensuring that each indicator has the maximum local gradient near its respective health threshold, significantly improving the system's sensitivity under critical conditions; employing different mapping functions such as Sigmoid product, Gaussian function, and Logistic to adapt to the interval sensitivity of illuminance, the symmetrical sensitivity of color temperature, and the unilateral sensitivity of blue light and glare, instead of uniformly using linear normalization; organically combining the normalized feature components with the gain / constraint coefficients output by the differentiated weight adjustment module 32 to form a weighted fusion framework, highlighting the impact of high-risk indicators while retaining the reference value of conventional indicators; and finally outputting feature data. This single, comprehensive indicator simplifies the input dimensions for subsequent compliance decision-making units, while retaining intermediate feature components to support multi-dimensional diagnostics. All normalized parameters ( All sensors can be calibrated online according to actual sensor characteristics and health standards, ensuring the system's engineering adaptability and accuracy.

[0085] The Healthy Lighting Compliance Decision Unit 4 receives high-precision lighting environment characteristic data, completes compliance verification through a built-in classroom healthy lighting threshold judgment model, and generates healthy lighting control instructions adapted to the teaching environment by combining real-time personnel distribution and environmental conditions in the classroom. Specifically, it includes an environmental characteristic compliance verification module 41, an operating condition information matching module 42, and a lighting instruction generation module 43, wherein: Environmental feature compliance verification module 41 is used to receive high-precision lighting environment feature data. The built-in classroom healthy lighting threshold judgment model is used to complete the compliance verification of lighting data; Specifically, the environmental feature compliance verification module 41 receives high-precision lighting environment feature data output by the adaptive fusion processing unit 3. Simultaneously receive each normalized feature component And the original measured values. The classroom healthy lighting threshold determination model built into the environmental feature compliance verification module 41 adopts a hierarchical threshold comparison and rule table structure.

[0086] The classroom healthy lighting threshold determination model first sets single-index physical quantity thresholds based on the national standards GB7793-2010 "Hygienic Standards for Daylighting and Illumination in Primary and Secondary School Classrooms" and GB / T36876-2018 "Hygienic Requirements for Lighting Design and Installation in Primary and Secondary School Classrooms": Illuminance The compliance range is 300 lx≤ ≤500 lx, the threshold for triggering regulation is <280 lx or >520lx (return difference 20 lx); Color temperature The compliance range is 4000 K≤ ≤5000 K; Blue light radiation intensity (415nm-455nm band) Compliance limit is ≤100 μW / cm 2 ; Unified Glare Index The compliance limit is ≤16.

[0087] The model outputs compliance indicators (compliant / non-compliant) for each indicator.

[0088] Furthermore, the classroom healthy lighting threshold determination model sets overall health characteristic thresholds. ,when The system determines overall compliance if the condition is met; otherwise, it determines non-compliance. Based on overall compliance and individual violations, the model outputs a 2-bit binary risk level code according to the following rules. : 00: Fully compliant ( And all individual items are compliant); 01: Poor comfort ( However, the illuminance or color temperature is not compliant, or (And both blue light and glare are compliant). 10: Health and safety risks ( (And blue light or glare is non-compliant). 11: Emergency comprehensive adjustment (any two or more non-compliant items and) ).

[0089] In addition, the model supports switching scene profiles (such as art class, evening self-study), and can adjust the compliance range of illuminance and color temperature; it also has a time-cumulative failure protection setting: when blue light or glare continues to be non-compliant for more than 10 minutes, even if... It also forced The threshold is increased to 10. All thresholds are configurable online. The verification results and risk codes are output to the lighting instruction generation module 43.

[0090] The working condition information matching module 42 is used to collect and match the real-time personnel distribution status in the classroom with the on-site environmental working conditions; specifically, the working condition information matching module 42 obtains the seat occupancy status of personnel through infrared human body sensors or seat pressure sensors deployed in the classroom, and outputs a set of occupied seats. and total number of people Outdoor illuminance is obtained through light sensors installed on the outdoor walls. The teaching type (regular lecture, multimedia teaching, examination) can be obtained through the course schedule interface or by manual input.

[0091] Furthermore, based on the above inputs, the working condition information matching module 42 calculates the following parameters: Key adjustment area marking: Limit lighting adjustments to areas where students are seated; Ambient light compensation coefficient ,in lx is the reference outdoor illuminance; Teaching type coefficient : Regular instruction is rated 1.0, multimedia teaching is rated 0.8, and examinations are rated 1.2.

[0092] The matching results are transmitted to the lighting instruction generation module 43 in the form of structured data.

[0093] The lighting instruction generation module 43 combines the compliance verification results with the real-time teaching environment conditions to generate healthy lighting control instructions adapted to the teaching environment. Specifically, the lighting instruction generation module 43 receives the risk level code output by the environmental characteristic compliance verification module 41. The system generates control instructions based on compliance indicators and the personnel distribution and environmental parameters output by the working condition information matching module 42, according to the priority principle of "blue light and glare safety take precedence over illuminance and color temperature comfort".

[0094] when When the value is 00, the system maintains the current lighting parameters or performs extremely fine adjustments. Target illuminance (Base target illuminance 400 lx corrected for environmental conditions) Illuminance adjustment is calculated using a proportional-integral controller, with the adjustment range limited to ±5% of the total light output.

[0095] when When the value is 01, a comfort adjustment command is generated: if the measured illuminance is below 300 lx, the drive duty cycle is increased; if it is above 500 lx, the duty cycle is decreased; if the measured color temperature deviates from 4500 K by more than 500 K, adjustment signals are sent to the cool white and warm white channels to gradually approach 4500 K.

[0096] when When the value is 10 or 11, a health and safety priority adjustment instruction is generated: If blue light exceeds the standard ( >100 μW / cm 2 The command reduces the PWM duty cycle of the blue light channel by 2% in each control cycle of the lighting fixture, while appropriately increasing the warm white light to maintain illuminance until the blue light complies with regulations. If glare exceeds the standard ( >16), instructing luminaires to reduce the intensity of direct light sources or to incorporate electronically controlled diffuser films to reduce... Seriously exceeding the standard ( When the light intensity is greater than 22), the direct-light downlight will be forcibly turned off and switched to indirect lighting.

[0097] During the safety adjustment process, if insufficient illuminance is caused, try to keep the illuminance no less than 200 lx and the color temperature between 3500 K and 6500 K.

[0098] The final output control command includes: target luminaire ID or area ID, adjustment type (illuminance / color temperature / blue light ratio / glare suppression), adjustment direction and step size or target value, and execution timestamp. It is encapsulated in binary protocol and sent to the smart lighting adaptive drive control unit 5.

[0099] The intelligent lighting adaptive drive control unit 5 receives healthy lighting control commands and performs high-frequency PWM stepless dimming and color adjustment operations on the classroom's whole-area lighting terminal equipment to achieve dynamic adjustment of lighting illuminance, color temperature, blue light ratio, and glare suppression across the entire area. Specifically, it includes a control command receiving and parsing module 51, a high-frequency PWM drive execution module 52, and a whole-area lighting dynamic adjustment module 53, wherein: The control instruction receiving and parsing module 51 is used to receive healthy lighting control instructions and complete instruction recognition and adaptation parsing; specifically, the control instruction receiving and parsing module 51 receives the control instructions sent by the healthy lighting compliance decision unit 4 through RS-485 or wireless communication interface.

[0100] The control instruction receiving and parsing module 51 performs CRC verification on the received data frame to filter out erroneous instructions; it also parses the target lamp ID or area ID in the instruction header to determine whether it falls within the jurisdiction of this unit.

[0101] For valid commands, the control command receiving and parsing module 51 converts the standardized command content into internal driving parameters: for example, "increase illuminance by 5%" is converted into a target duty cycle increase coefficient of 0.05 for the corresponding channel; "set blue light ratio to 0.3" is converted into a target value of 0.3 times the rated value of the blue LED channel PWM duty cycle.

[0102] If multiple commands for the same lamp are received within a short period, the last command will be used to overwrite the previous one. The parsed drive parameters are stored in the command queue, waiting to be read by the high-frequency PWM drive execution module 52.

[0103] The high-frequency PWM drive execution module 52 outputs a high-frequency PWM stepless drive signal based on the parsed control instructions. Specifically, the high-frequency PWM drive execution module 52 sets the PWM carrier frequency to 30 kHz, which is much higher than the critical flicker frequency of the human eye, thus eliminating flicker. The PWM duty cycle adjustment resolution reaches 12 bits (4096 levels), achieving stepless smooth dimming. The high-frequency PWM drive execution module 52 supports independent control of multiple channels, generating independent PWM signals for dedicated channels for cool white, warm white, and blue light. The duty cycle change adopts an S-shaped acceleration curve, and the transition time is automatically adjusted between 50 ms and 500 ms according to the adjustment amplitude to avoid visual discomfort caused by sudden changes. The high-frequency PWM drive execution module 52 integrates an LED drive current feedback loop to monitor the actual current of each channel in real time and perform closed-loop calibration with the expected current corresponding to the target duty cycle, eliminating nonlinear errors caused by temperature and aging.

[0104] The omnidirectional lighting dynamic adjustment module 53 relies on a high-frequency PWM drive signal to dynamically adjust the illuminance, color temperature, blue light ratio, and glare suppression of the classroom omnidirectional lighting terminal equipment.

[0105] Specifically, the global lighting dynamic adjustment module 53 performs the following adjustments based on the drive signal output by the high-frequency PWM drive execution module 52: Illuminance adjustment: This is achieved by simultaneously adjusting the PWM duty cycle of the cool white and warm white channels (keeping their ratio constant), with an adjustment range of 50 lx to 800 lx. When the ambient light compensation coefficient is low, the maximum output illuminance is automatically reduced to save energy.

[0106] Color temperature adjustment: This is achieved by changing the PWM duty cycle ratio between the cool white channel and the warm white channel, with an adjustment range of 3000K to 6500K. Since the LED-related color temperature and duty cycle have a non-linear relationship and are affected by individual LED differences and temperature drift, this embodiment adopts an engineering implementation method using a pre-calibrated duty cycle-color temperature reference table. Before leaving the factory, the system uses a spectrophotometer to measure the actual correlated color temperature corresponding to different combinations of cool white / warm white duty cycles, establishes a high-precision lookup table (LUT), and stores it in the controller's non-volatile memory. The lookup table covers the full adjustment range from 3000K to 6500K, with a color temperature step size of ≤50K, ensuring a mixed light color temperature error of ≤±100K; For target color temperatures not included in the lookup table, the corresponding duty cycle combination is calculated using linear interpolation.

[0107] The core parameters involved in the above adjustments are defined as follows: This represents the PWM duty cycle of the cold white channel, which is dimensionless and ranges from [0,1]. This indicates the rated color temperature of the cool white channel, which is 6500 K in this embodiment; This represents the PWM duty cycle of the warm white channel, dimensionless, with a value range of [0,1]. This indicates the rated color temperature of the warm white channel, which is 3000 K in this embodiment.

[0108] The system calculates the required cool white / warm white duty cycle ratio based on the target color temperature using query or interpolation, and simultaneously fine-tunes the total drive intensity (maintaining...). + (Basically constant) to ensure that the illuminance remains basically unchanged.

[0109] Blue light ratio adjustment: For lamps equipped with independent blue LED channels (peak wavelength 450nm), the PWM duty cycle of the blue light channel is controlled separately. This allows for precise adjustment of blue light radiation intensity. The blue light ratio is defined as: ; in: This represents the proportion of blue light radiation power to total radiation power, and is dimensionless. This represents the PWM duty cycle of the Blu-ray channel, which is dimensionless and ranges from [0,1]. These are the radiant fluxes (in W) of the cool white, warm white, and blue light channels at rated current, respectively.

[0110] When receiving instructions to reduce blue light risk, gradually reduce Until it approaches 0, while increasing the duty cycle of the warm white channel to compensate for the illuminance loss.

[0111] Glare suppression is achieved in two ways. First, by reducing direct light intensity: the drive duty cycle of direct-light fixtures such as blackboard lights and downlights is reduced according to instructions, or the luminous flux is transferred to ceiling reflection. Second, by switching diffusion optical elements: for fixtures equipped with electronically controlled diffusion films or adjustable shading angle mechanisms, the output control signal cuts the diffusion film into the light path or increases the shading angle to reduce the uniform glare index (UGR).

[0112] Independent zone adjustment: Based on the personnel distribution information provided by the working condition information matching module 42, fine adjustment is performed only in occupied areas, while unoccupied areas maintain the minimum safe illuminance (50 lx) or turn off the lights to achieve energy saving.

[0113] Finally, the global lighting dynamic adjustment module 53 periodically transmits all adjustment statuses (current illuminance, color temperature, blue light ratio, PWM duty cycle of each channel, etc.) back to the healthy lighting compliance decision-making unit 4, forming a closed-loop feedback.

[0114] Those skilled in the art will understand that the process of implementing all or part of the steps of the above embodiments can be carried out by hardware or by a program instructing the relevant hardware.

[0115] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the claimed invention.

Claims

1. A smart classroom health lighting system based on multi-sensor fusion, characterized in that, include: Multi-source sensor synchronous acquisition unit (1) acquires various types of raw environmental data in the smart classroom, performs synchronous processing and timestamp marking, and generates raw sensor dataset with time sequence identifier. Health risk analysis unit (2) receives the original sensor dataset and performs time registration and coarse noise reduction processing in sequence. Based on the preset health thresholds of various sensor indicators, it constructs an instantaneous health risk evaluation function, calculates the time cumulative risk factors corresponding to blue light and glare indicators, couples instantaneous health risk and time cumulative risk factors to obtain comprehensive health risk, and completes the risk stratification of lighting indicators based on comprehensive health risk. The adaptive fusion processing unit (3) receives the risk stratification results, carries heterogeneous data priority fusion logic, enables the health threshold neighborhood sensitivity adjustment mechanism, increases the sampling frequency and weight gain slope of blue light and glare high-risk indicators in the critical health state, applies weight gain adjustment to blue light and glare high-risk health indicators, applies weight constraint adjustment to light illuminance conventional comfort indicators, completes data fusion through differentiated normalization processing that retains the health threshold boundary characteristics, and outputs high-precision lighting environment feature data. The healthy lighting compliance decision unit (4) receives high-precision lighting environment characteristic data, completes compliance verification through the built-in classroom healthy lighting threshold judgment model, and generates healthy lighting control instructions adapted to the teaching environment by combining the real-time personnel distribution and environmental conditions in the classroom. The intelligent lighting adaptive drive control unit (5) receives the health lighting control command and performs high-frequency PWM stepless dimming and color adjustment operation on the classroom whole-area lighting terminal equipment to realize the whole-area dynamic adjustment of lighting illuminance, color temperature, blue light ratio and glare suppression.

2. The smart classroom health lighting system based on multi-sensor fusion according to claim 1, characterized in that, The multi-source sensor synchronous acquisition unit (1) includes a sensor acquisition module (11), a data synchronization processing module (12), and a time sequence marker generation module (13), wherein: The sensor acquisition module (11) collects various types of raw environmental data in the smart classroom; The data synchronization processing module (12) performs synchronization processing on multiple types of raw environmental data collected by the sensor acquisition module (11). The time stamp generation module (13) timestamps the data after it has been synchronized by the data synchronization processing module (12) to generate the original sensor dataset with time stamp.

3. The smart classroom health lighting system based on multi-sensor fusion according to claim 2, characterized in that, The health risk analysis unit (2) includes a preprocessing and instantaneous risk calculation module (21); the preprocessing and instantaneous risk calculation module (21) is used to purify and preprocess the original sensor dataset and quantify the instantaneous health risks of various lighting indicators; the data preprocessing and instantaneous risk calculation process of the preprocessing and instantaneous risk calculation module (21) includes the following steps: S21.1 Receive the raw sensor dataset, perform time registration processing and coarse noise reduction processing in sequence, unify the time reference of all sensor data, and remove abnormal data in the dataset that exceed the physical range. S21.2 Construct an instantaneous health risk assessment function based on preset health thresholds for various sensor indicators, and calculate and output the instantaneous health risk values ​​corresponding to various sensor indicators. .

4. The smart classroom health lighting system based on multi-sensor fusion according to claim 3, characterized in that, The health risk analysis unit (2) further includes a time-cumulative risk calculation module (22); the time-cumulative risk calculation module (22) is used to quantify the time-series cumulative risk based on instantaneous health risk for high-risk lighting indicators; the time-cumulative risk factor calculation process of the time-cumulative risk calculation module (22) includes the following steps: S22.1 Receive the instantaneous health risk value output by the preprocessing and instantaneous risk calculation module (21). From this, values ​​corresponding to two categories of indicators, blue light and glare, are selected. S22.2 Activate the preset sliding time window statistical mode, combining the time series data within the window with the instantaneous health risk value. Calculate and output the time-cumulative risk factor of the corresponding indicator. .

5. The smart classroom health lighting system based on multi-sensor fusion according to claim 4, characterized in that, The health risk analysis unit (2) also includes a comprehensive risk coupling and stratification module (23); the comprehensive risk coupling and stratification module (23) is used to integrate instantaneous risk and cumulative risk, and to complete the classification of lighting health risk levels; the risk coupling and indicator stratification process of the comprehensive risk coupling and stratification module (23) includes the following steps: S23.1 Receive instantaneous health risk value With time-cumulative risk factors The comprehensive health risk value of each lighting indicator is obtained by coupling through a weighted fusion method. ; S23.2, Based on the calculated comprehensive health risk value The risk stratification of lighting indicators was completed, classifying blue light and glare indicators as high-risk health indicators, and illuminance indicators as routine comfort indicators.

6. The smart classroom health lighting system based on multi-sensor fusion according to claim 5, characterized in that, The adaptive fusion processing unit (3) includes a risk reception and neighborhood adjustment module (31); the risk reception and neighborhood adjustment module (31) is used to receive risk stratification results and perform threshold neighborhood sensitivity and sampling parameter adjustment; the data reception and parameter adjustment process of the risk reception and neighborhood adjustment module (31) includes the following steps: S31.1 Receive the risk stratification results output by the integrated risk coupling and stratification module (23) and enable the heterogeneous data priority fusion logic; S31.2 Activate the health threshold neighborhood sensitivity adjustment mechanism to increase the sampling frequency and weight gain slope of blue light and glare high-risk indicators in the critical health state. .

7. The smart classroom health lighting system based on multi-sensor fusion according to claim 6, characterized in that, The adaptive fusion processing unit (3) further includes a differential weight adjustment module (32); the differential weight adjustment module (32) is used to set weight rules for different types of lighting indicators; the indicator weight adjustment process of the differential weight adjustment module (32) includes the following steps: S32.1 Apply weighted gain adjustment to high-risk health indicators such as blue light and glare, and configure the weighted gain coefficient. ; S32.2 Apply weighted constraints to the conventional comfort index of illuminance and configure the weighted constraint coefficients. .

8. The smart classroom health lighting system based on multi-sensor fusion according to claim 7, characterized in that, The adaptive fusion processing unit (3) further includes a differential normalization fusion module (33); the differential normalization fusion module (33) is used to complete data normalization, multi-indicator fusion and output feature data; The data normalization and fusion output process of the differential normalization fusion module (33) includes the following steps: S33.

1. Differential normalization processing is completed by preserving the boundary characteristics of the health threshold, thereby achieving the fusion of multiple types of indicator data; S33.2 Output the high-precision lighting environment feature data obtained by fusion. .

9. The smart classroom health lighting system based on multi-sensor fusion according to claim 8, characterized in that, The healthy lighting compliance decision-making unit (4) includes an environmental characteristic compliance verification module (41), an operating condition information matching module (42), and a lighting instruction generation module (43), wherein: The environmental feature compliance verification module (41) is used to receive high-precision lighting environment feature data. The built-in classroom healthy lighting threshold judgment model is used to complete the compliance verification of lighting data; The working condition information matching module (42) is used to collect and match the real-time personnel distribution status in the classroom with the on-site environmental working conditions; The lighting instruction generation module (43) is used to combine the compliance verification results with the real-time teaching environment conditions to generate healthy lighting control instructions that are adapted to the teaching environment.

10. The smart classroom health lighting system based on multi-sensor fusion according to claim 9, characterized in that, The intelligent lighting adaptive drive control unit (5) includes a control command receiving and parsing module (51), a high-frequency PWM drive execution module (52), and a global lighting dynamic adjustment module (53), wherein: The control instruction receiving and parsing module (51) is used to receive health lighting control instructions and complete instruction recognition and adaptation parsing; The high-frequency PWM drive execution module (52) outputs a high-frequency PWM stepless drive signal based on the parsed control instructions; The global lighting dynamic adjustment module (53) relies on the high-frequency PWM drive signal to dynamically adjust the illuminance, color temperature, blue light ratio and glare suppression of the classroom global lighting terminal equipment.

Images