LED color-emotion association map construction method and intelligent lighting control system

Abstract

The invention belongs to the technical field of LED illumination, and particularly relates to an LED color-emotion association map construction method and an intelligent illumination control system, and the method comprises the steps: constructing a multi-dimensional emotion evaluation system; determining full-dimensional color parameters of the LED light source; designing a control variable method experiment and collecting multi-modal data, wherein the experiment is carried out in a standardized test environment; processing multi-modal data and establishing a mapping relation; and constructing a three-dimensional association map. According to the LED color-emotion association map construction method and the intelligent illumination control system, emotion adjustment precision improvement, individual suitability optimization, dynamic response capability enhancement and product application value improvement can be realized.

Description

Technical Field

[0001] This invention belongs to the field of LED lighting technology, and in particular relates to a method for constructing an LED color-emotion association map and an intelligent lighting control system. Background Technology

[0002] I. Current Status of Technology.

[0003] Current LED lighting products have been upgraded from "basic lighting" to "scenario-based lighting", but color adjustment technology still has significant limitations.

[0004] 1. Lack of scientific basis for color-emotion matching: Existing products mostly use "empirical color temperature zoning" (such as 2700K = warm light = relaxation), without establishing a quantitative mapping between color parameters (such as saturation gradient, spectral peak) and emotional dimensions (such as arousal, dominance), resulting in insufficient precision in emotion regulation. For example, existing "sleep aid lamps" only have a fixed color temperature of 2700K, without considering the differences in sensitivity to saturation among different groups of people. In actual tests, 30% of users reported that "warm light is still dazzling".

[0005] 2. Lack of individual variation adaptation: The existing system does not integrate individual variables such as age, color vision characteristics, and cultural background.

[0006] Age differences: Due to the yellowing of the lens, older adults have a reduced perception of 450-500nm blue light and need 15%-20% more blue light than younger people to achieve the same "clear" mood.

[0007] Cultural differences: In Western markets, "red" equals "warning," while in Eastern markets, "red" equals "festive." The uniform use of red parameters for existing products has led to a difference of over 40% in user satisfaction across different regions.

[0008] Color vision differences: Approximately 8% of men have red-green color weakness. Existing high-saturation red / green parameters can easily cause visual fatigue in color-weak users, but no compensation color scheme has been set.

[0009] 3. Limited data collection dimensions: Most solutions rely solely on subjective emotion scores without combining physiological signals (such as cortisol levels and heart rate variability) and neural electrical signals (such as EEG brainwaves), resulting in subjective bias in "emotional feedback." Experimental data shows that, under the same color parameter, when a user's subjective score is "pleasure = 6 points," the skin conductivity in the physiological signal may fluctuate by ±2μS, reflecting a difference between the actual emotional state and the subjective expression.

[0010] 4. Weak dynamic adjustment mechanism: Existing systems are mostly "one-time parameter matching", which cannot cope with dynamic changes in ambient light (such as cloudy days vs. sunny days) and fluctuations in user emotions (such as work fatigue vs. relaxation). For example, in office scenarios, users are prone to a decline in attention between 2:00 pm and 4:00 pm. The existing fixed 5000K color temperature cannot be adjusted in real time, resulting in a 15%-20% reduction in work efficiency.

[0011] II. Reasons for the technical pain points.

[0012] 1. Insufficient interdisciplinary technology integration: The lack of technological collaboration between the field of LED engineering and the fields of emotional psychology and biosensing has resulted in the design of color parameters not taking into account the "neural mechanism of emotional perception" (such as the amygdala's arousal response to red).

[0013] 2. Inadequate experimental design: Existing correlation studies often employ "small sample uncontrolled experiments" (e.g., sample size <50 people, uncontrolled ambient light interference), resulting in low data reliability and an inability to support large-scale product applications.

[0014] 3. Lack of hardware-algorithm synergy: Most systems only focus on adjusting software parameters and do not optimize LED hardware characteristics (such as spectral continuity and color deviation stability), resulting in a deviation of DeltaE=3-5 between "theoretical parameters" and "actual lighting effect", which is beyond the acceptable range of the human eye (DeltaE≤2).

[0015] Therefore, it is urgent to address the technical deficiencies of existing technologies through multimodal data fusion, standardized experimental design, and hardware-algorithm collaboration. Summary of the Invention

[0016] The purpose of this invention is to provide a method for constructing an LED color-emotion correlation map and an intelligent lighting control system. The aim is to construct a three-dimensional correlation model of "color parameters-emotional dimension-physiological indicators" to realize the quantitative definition of the emotional effect of LED colors, so as to solve the problem of emotional regulation adaptability caused by individual differences (age, culture, color vision), and finally realize closed-loop control of "environmental perception-emotion recognition-parameter adjustment-feedback optimization" to improve the emotional adaptation accuracy and dynamic response capability of LED lighting.

[0017] To achieve the above objectives, this invention provides a method for constructing an LED color-emotion association map, comprising the following steps:

[0018] S1: Construct a multi-dimensional emotion assessment system, which includes core emotion dimensions and derived indicators. The core emotion dimensions include pleasure, arousal, fit, and dominance. The derived indicators include emotion duration and emotion triggering delay time.

[0019] The core emotional dimensions are quantified using a 1-7 point PAD emotional scale. The pleasantness is quantified with the help of the Visual Analogue Scale (VAS). The arousal level records the emotional trigger delay time in sync. The suitability is weighted in combination with the functional requirements of the scene. The advantage level is related to the parameter adjustment of the willingness index.

[0020] S2: Determine the full-dimensional color parameters of the LED light source, including color temperature, saturation, brightness, color gamut type, color deviation value, dynamic gradient speed, and spectral peak wavelength;

[0021] The color temperature range is 2200K-6500K, with an adjustment accuracy of ±50K;

[0022] The saturation range is 0%-100%, with an adjustment accuracy of ±1%.

[0023] The brightness is divided into lighting scenarios (50 lux-800 lux, accuracy ±5 lux) and display scenarios (50 nit-500 nit, accuracy ±2 nit).

[0024] The color gamut types include sRGB, DCI-P3, and AdobeRGB, with a coverage of ≥98%;

[0025] The color deviation values ​​include general-grade DeltaE≤3 and professional-grade DeltaE≤1, with an adjustment accuracy of ±0.2.

[0026] The range of the dynamic gradual speed change is 0.5s-5s, with an accuracy of ±0.1s;

[0027] The peak wavelength of the spectrum is in the range of 400nm-760nm, with an accuracy of ±5nm;

[0028] S3: Design an experiment using the controlled variable method and collect multimodal data. The experiment should be conducted in a standardized testing environment.

[0029] The standardized testing environment meets the following requirements: spatial dimensions of 3m × 3m × 2.8m, with matte neutral gray inner walls. Ambient illuminance ≤50 lux, D50 standard light source color temperature 5000K, humidity 40%-60%, temperature 23±2℃;

[0030] A total of 300 participants were divided into 6 age groups, with 50 participants aged 18-25 and 50 participants aged 56-60, half male and half female, covering 4 cultural backgrounds: East, West, Southeast Asia and Middle East, and all of them passed the Ishihara color vision test.

[0031] The experiment uses a random sequence to present color parameter combinations, ensuring that each combination is presented for 10 seconds with a 3-second interval, and that data is collected synchronously.

[0032] Subjective rating, which is a 10-second timed rating based on the four dimensions of PAD;

[0033] Physiological signals, namely skin conductivity sampled at 10 Hz, heart rate sampled at 5 Hz, and salivary cortisol sampled once before and once after the test;

[0034] The neural electrical signals were sampled at 250Hz using a 64-channel EEG cap, with a focus on acquiring the F3 / F4, T3 / T4, and O1 / O2 regions.

[0035] Each participant completed 3 rounds of testing, with each round spaced 24 hours apart. The third round included an additional parameter: memory assessment.

[0036] S4: Multimodal data processing and establishing mapping relationships, including:

[0037] Outlier removal, the outlier removal is as follows: Criteria plus manual verification of logically contradictory data;

[0038] Data standardization, which involves Z-score processing of physiological signals and 0.5-30Hz bandpass filtering + ICA artifact removal of EEG signals;

[0039] Data alignment, wherein the data alignment synchronizes three types of data according to timestamps;

[0040] The Pearson correlation coefficient method was used to analyze the association between parameters and emotions. A strong association was defined as r ≥ 0.7, P < 0.001. A multiple linear regression model was constructed, and physiological index thresholds were calibrated. The goodness of fit of the multiple linear regression model was then analyzed. The physiological indicator thresholds include relaxation and sleep aid, corresponding to skin conductivity ≤3μS, heart rate ≤70 beats / minute, and EEG α band amplitude ≥8μV;

[0041] S5: Construct a three-dimensional correlation map. The three-dimensional correlation map uses the "core color parameter combination" composed of color temperature, saturation, and brightness as the X-axis, the emotional dimension quantification composed of pleasure, arousal, fit, and dominance as the Y-axis, and the physiological signal threshold range composed of physiological signals as the Z-axis. It is divided into home layer, office layer, commercial layer, and medical layer according to the scene. Each mapping point is marked with a 95% confidence interval. The confidence level <90% is marked "to be verified". The map supports multi-dimensional retrieval of emotional goals, scenes, and physiological signals, and can automatically correct parameters by inputting user characteristics.

[0042] This invention also provides an intelligent lighting control system, which uses the three-dimensional correlation map obtained by the above-described LED color-emotion correlation map construction method, and includes:

[0043] The data acquisition module includes a scene recognition unit, an ambient light detection unit, and a physiological signal acquisition unit;

[0044] The scene recognition unit uses a 2-megapixel RGB camera and an infrared sensor (0.5-5m detection) to recognize 6 types of scenes, including bedrooms and offices, through the YOLOv5+ scene classification model, and is automatically calibrated every 7 days;

[0045] The ambient light detection unit uses an AMSTCS34725 spectral sensor and a VEML7700 illuminance sensor, with detection errors of color temperature ±100K and illuminance ±5%, sampling at 1Hz and constructing an ambient light change curve.

[0046] The physiological signal acquisition unit supports Bluetooth 5.0 connection to wearable devices (10Hz synchronous sampling) or integrates an Ag / AgCl patch sensor, and automatically switches to the default scene parameters when the signal is interrupted.

[0047] The associated map storage module includes dual storage: local and cloud storage. The local storage contains 50MB of basic map data, while the cloud storage contains 200MB of complete map data. Local data is encrypted with AES-256, and cloud data complies with the ISO27001 standard, ensuring user data anonymization.

[0048] The emotion needs matching module has a built-in scene-emotion mapping library and uses a multi-factor matching algorithm: basic matching, individual correction, environmental adaptation, and conflict handling.

[0049] The lighting control module includes an LED driver unit and a color management unit;

[0050] The LED driver unit uses an STM32F407MCU+TITPS92661 four-channel driver chip, with constant current 0-1A output, adjustment error of color temperature ±50K, saturation ±1%, brightness ±5lux, response ≤100ms, and adopts PWM and linear hybrid dimming.

[0051] The color management unit supports Conversion, spectral calibration every 24 hours; supports 16-zone linkage;

[0052] The dynamic optimization module collects both active and passive feedback; it uses gradient descent for short-term optimization and random forest algorithm for long-term optimization, with the population iterating over a general graph in the cloud every quarter.

[0053] The user interaction module supports voice and APP interaction, and is used to generate user profiles and recommend solutions. It supports 50 sets of solutions to be collected and shared, or setting a timer for the emotion curve.

[0054] The fault self-diagnosis module is used to monitor the drive current, voltage, and spectrum sensors in real time to determine hardware faults, or to monitor parameter deviations to determine parameter out of control; for first-level faults, it switches to backup parameters and provides a prompt; for second-level faults, it shuts down the fault zone, starts backup lighting, and generates a fault code; 100 fault records can be stored locally and exported.

[0055] The LED color-emotion association map construction method and intelligent lighting control system provided in this invention have at least one of the following technical effects: improved emotion regulation accuracy; the LED system using this invention achieves an emotion target matching accuracy of 89%, which is 43.5% higher than the traditional fixed parameter system (62%); physiological indicator verification: under the "sleep aid mode", the average time for users to fall asleep is shortened by 18 minutes, and the number of awakenings at night is reduced by 2.1 times / night (PSQI sleep quality index drops from 12.3 to 7.8). Optimized individual adaptability; for users aged 56-60, the emotional satisfaction rate after parameter correction reaches 87%, which is 33.8% higher than the uncorrected solution (65%); the satisfaction rate of Eastern users with red parameters reaches 90%, and the satisfaction rate of Western users with red parameters reaches 85% (compared to 72% and 60% for the unadapted solutions, respectively). Enhanced dynamic response capability; Environmental adaptation: When ambient light increases from 50 lux (cloudy day) to 500 lux (sunny day), the system parameter adjustment response time is ≤500ms, and emotional satisfaction remains above 85% (compared to 68% for traditional systems); Emotional fluctuation adaptation: When the user's heart rate increases from 70 beats / minute to 90 beats / minute (anxiety), the system automatically switches to "soothing mode," and the heart rate decreases by an average of 12 beats / minute after 5 minutes. Increased product application value; Office scenario: After adopting this system, user work efficiency increases by 19% (task completion time is reduced by 22 minutes / day), and the Visual Acuity Scale (VAS) score decreases from 6.2 to 3.1; Medical scenario: After application in hospital wards, the patient anxiety scale (GAD-7) score decreases from 8.5 to 5.2, and the number of nurses' nighttime calls decreases by 35%. Attached Figure Description

[0056] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0057] Figure 1 A flowchart illustrating the method for constructing an LED color-emotion association map according to an embodiment of the present invention.

[0058] Figure 2 This is a structural block diagram of an intelligent lighting control system provided in an embodiment of the present invention. Detailed Implementation

[0059] The following is in conjunction with the appendix Figures 1-2 The embodiments of the present invention are described in detail below. These embodiments are exemplary and intended to explain the embodiments of the present invention, and should not be construed as limiting the present invention.

[0060] In one embodiment of the present invention, such as Figure 1 As shown, a method for constructing an LED color-emotion association graph is provided, applicable to LED lighting in various scenarios such as home, office, commercial, and medical settings. The method includes the following steps:

[0061] S1: Construct a multi-dimensional emotion assessment system, which includes core emotion dimensions and derived indicators. The core emotion dimensions include pleasure, arousal, fit, and dominance. The derived indicators include emotion duration and emotion triggering delay time.

[0062] The core emotional dimensions are quantified using a 1-7 point PAD emotional scale. The pleasantness is quantified with the help of the Visual Analogue Scale (VAS). The arousal level records the emotional trigger delay time in sync. The suitability is weighted in combination with the functional requirements of the scene. The advantage level is related to the parameter adjustment of the willingness index.

[0063] S2: Determine the full-dimensional color parameters of the LED light source, including color temperature, saturation, brightness, color gamut type, color deviation value, dynamic gradient speed, and spectral peak wavelength;

[0064] The color temperature range is 2200K-6500K, with an adjustment accuracy of ±50K;

[0065] The saturation range is 0%-100%, with an adjustment accuracy of ±1%.

[0066] The brightness is divided into lighting scenarios (50 lux-800 lux, accuracy ±5 lux) and display scenarios (50 nit-500 nit, accuracy ±2 nit).

[0067] The color gamut types include sRGB, DCI-P3, and AdobeRGB, with a coverage of ≥98%;

[0068] The color deviation values ​​include general-grade DeltaE≤3 and professional-grade DeltaE≤1, with an adjustment accuracy of ±0.2.

[0069] The range of the dynamic gradual speed change is 0.5s-5s, with an accuracy of ±0.1s;

[0070] The peak wavelength of the spectrum is in the range of 400nm-760nm, with an accuracy of ±5nm;

[0071] S3: Design an experiment using the controlled variable method and collect multimodal data. The experiment should be conducted in a standardized testing environment.

[0072] The standardized testing environment meets the following requirements: spatial dimensions of 3m × 3m × 2.8m, with matte neutral gray inner walls. Ambient illuminance ≤50 lux, D50 standard light source color temperature 5000K, humidity 40%-60%, temperature 23±2℃;

[0073] A total of 300 participants were divided into 6 age groups, with 50 participants aged 18-25 and 50 participants aged 56-60, half male and half female, covering 4 cultural backgrounds: East, West, Southeast Asia and Middle East, and all of them passed the Ishihara color vision test.

[0074] The experiment uses a random sequence to present color parameter combinations, ensuring that each combination is presented for 10 seconds with a 3-second interval, and that data is collected synchronously.

[0075] Subjective rating, which is a 10-second timed rating based on the four dimensions of PAD;

[0076] Physiological signals, namely skin conductivity sampled at 10 Hz, heart rate sampled at 5 Hz, and salivary cortisol sampled once before and once after the test;

[0077] The neural electrical signals were sampled at 250Hz using a 64-channel EEG cap, with a focus on acquiring the F3 / F4, T3 / T4, and O1 / O2 regions.

[0078] Each participant completed 3 rounds of testing, with each round spaced 24 hours apart. The third round included an additional parameter: memory assessment.

[0079] S4: Multimodal data processing and establishing mapping relationships, including:

[0080] Outlier removal, the outlier removal is as follows: Criteria plus manual verification of logically contradictory data;

[0081] Data standardization, which involves Z-score processing of physiological signals and 0.5-30Hz bandpass filtering + ICA artifact removal of EEG signals;

[0082] Data alignment, wherein the data alignment synchronizes three types of data according to timestamps;

[0083] The Pearson correlation coefficient method was used to analyze the association between parameters and emotions. A strong association was defined as r ≥ 0.7, P < 0.001. A multiple linear regression model was constructed, and physiological index thresholds were calibrated. The goodness of fit of the multiple linear regression model was then analyzed. The physiological indicator thresholds include relaxation and sleep aid, corresponding to skin conductivity ≤3μS, heart rate ≤70 beats / minute, and EEG α band amplitude ≥8μV;

[0084] S5: Construct a three-dimensional correlation map. The three-dimensional correlation map uses the "core color parameter combination" composed of color temperature, saturation, and brightness as the X-axis (encoding format such as 2700K_50%_60lux), the emotional dimension quantification composed of pleasure, arousal, fit, and dominance as the Y-axis (including derived indicators), and the physiological signal threshold range composed of physiological signals as the Z-axis. It is divided into home layer, office layer, commercial layer, and medical layer according to the scene (each layer is labeled with emotional target weight). Each mapping point is labeled with a 95% confidence interval, and the confidence level <90% is labeled "to be verified". The map supports multi-dimensional retrieval of emotional targets, scenes, and physiological signals, and can automatically correct parameters by inputting user characteristics (age, culture, color vision).

[0085] Furthermore, in step S3, the subject needs to undergo 10 minutes of environmental adaptation by watching a neutral gray screen before the test, and basic physiological signals are collected simultaneously as a baseline; during EEG signal analysis, the amplitude and latency data of the P1 band (80-120ms), N1 band (150-200ms), and LPP band (300-600ms) are extracted.

[0086] Further, in step S4, the formula of the multiple linear regression model is: Pleasure (P) = a1 × color temperature + a2 × saturation + a3 × brightness + a4 × spectral peak value + b, where a1-a4 are regression coefficients (e.g., a1 = -0.002, which represents a decrease of 0.2 points in pleasure for every 100K increase in color temperature), and b is a constant term.

[0087] like Figure 2 As shown, this embodiment of the invention also provides an intelligent lighting control system, which uses the three-dimensional correlation map obtained by the above-described LED color-emotion correlation map construction method, and includes:

[0088] The data acquisition module includes a scene recognition unit, an ambient light detection unit, and a physiological signal acquisition unit;

[0089] The scene recognition unit uses a 2-megapixel RGB camera (1080P@30fps) and an infrared sensor (0.5-5m detection). It uses the YOLOv5+ scene classification model to recognize 6 types of scenes, such as bedroom and office (accuracy ≥95%, latency ≤300ms), and automatically calibrates every 7 days.

[0090] The ambient light detection unit uses an AMSTCS34725 spectral sensor (400-700nm) + VEML7700 illuminance sensor (0-1000lux), with detection errors of color temperature ±100K and illuminance ±5%, sampling at 1Hz and constructing an ambient light change curve.

[0091] The physiological signal acquisition unit supports Bluetooth 5.0 connection to wearable devices (10Hz synchronous sampling), or integrates an Ag / AgCl patch sensor (impedance). USB-C powered, latency ≤100ms), automatically switches to default scene parameters when signal is interrupted;

[0092] The associated map storage module includes dual local and cloud storage. The local storage contains 50MB of basic map data (usable offline), while the cloud storage contains 200MB of complete map data (including iterative data). Local data is encrypted with AES-256, while cloud data complies with ISO27001 standards, and user data is anonymized. Maps are managed by "V year. quarter" version and support rolling back to historical versions.

[0093] The emotional needs matching module has a built-in scene-emotion mapping library (e.g., the core emotion in the bedroom at night is "relaxation and sleep aid", with an arousal weight of 0.6 and a priority level of 1). It adopts a multi-factor matching algorithm: ① basic matching (parameters are extracted according to the scene), ② individual correction (e.g., color temperature +100K and saturation +5% for a 56-year-old Asian user), ③ environmental adaptation (correction based on a standard environment of 5000K / 50lux), and ④ conflict handling (fusion of multiple objectives according to priority or parameters).

[0094] The lighting control module includes an LED driver unit and a color management unit;

[0095] The LED driving unit uses an STM32F407MCU + TITPS92661 four-channel driving chip, with constant current output of 0-1A, adjustment error of color temperature ±50K, saturation ±1%, brightness ±5lux, response ≤100ms, and adopts PWM (high brightness) + linear (low brightness <100lux) hybrid dimming;

[0096] The color management unit supports Conversion (error ≤ ΔE 0.5), spectral calibration every 24 hours (adjust RGBW current if deviation > 10%); supports 16-zone linkage (master-slave partition latency ≤ 500ms).

[0097] The dynamic optimization module collects active feedback (1-5 points of satisfaction + text feedback) and passive feedback (e.g., if the heart rate in focus mode is >85 beats / minute, it is considered unsuitable); in the short term, gradient descent is used (the color temperature is reduced by 100K if the feedback score is 3 points), and in the long term, random forest algorithm is used (the individual model is updated weekly), and the general map of the group is iterated in the cloud every quarter.

[0098] The user interaction module supports voice (Mandarin / English / Cantonese, recognition rate ≥92%, 10 offline commands) and APP (professional mode slider adjustment / easy mode one-click switching) interaction, and is used to generate user profiles and recommend solutions. It supports 50 sets of solutions to collect and share, or setting mood curve timers (such as sleep aid mode 1 hour color temperature 2700K→2500K, brightness 80lux→50lux).

[0099] The fault self-diagnosis module is used to monitor the drive current (over 0-1A), voltage (fluctuation > ±10%), and spectral sensor (5 failed acquisitions) in real time to determine hardware faults, or to monitor parameter deviations (color temperature > 150K, saturation > 10%, brightness > 20 lux for 10 seconds) to determine parameter out-of-control. For first-level faults, it switches to backup parameters (such as 2700K / 50% / 100lux) and provides a prompt. For second-level faults, it shuts down the fault zone, starts backup lighting, and generates a fault code (such as ERR-DRV-01). 100 fault records can be stored locally and exported.

[0100] Furthermore, the scenario-emotion mapping library includes the emotional goals and weight settings for retail store windows (core emotion "attracting attention", arousal weight 0.7, priority level 1), hospital wards (core emotion "soothing and calming", pleasure weight 0.5, priority level 1), and living room gatherings (core emotion "pleasant interaction", dominance weight 0.4, priority level 2).

[0101] Furthermore, the emotional needs matching module includes a cultural difference adaptation unit, in which the correction coefficient for the red series parameters of Eastern culture is 1.1-1.2, the correction coefficient for the red series parameters of Western culture is 0.8-0.9, the correction coefficient for the red series parameters of Southeast Asian culture is 0.95-1.05, and the correction coefficient for the red series parameters of Middle Eastern culture is 1.0-1.1.

[0102] Furthermore, the lighting control module can support multi-zone linkage control with two trigger modes, including timed triggering (switching parameters according to a preset time) and scene linkage triggering (such as when the main bedroom light switches to sleep aid mode, the ambient light simultaneously drops to 2700K / 40% / 30lux).

[0103] Furthermore, the iteration step size of the dynamic optimization module decreases as the number of feedbacks increases. When the number of feedbacks for the same parameter combination is ≥10 and the satisfaction score is ≥4.5, the parameter combination is set as the default parameter for the corresponding scenario.

[0104] Furthermore, the user interaction module's APP supports a family sharing function, which can add at least 5 family member accounts and is used to store personalized plans and support automatic switching of corresponding plans according to members' schedules.

[0105] Furthermore, the backup parameter schemes of the fault self-test module are pre-stored according to scenario categories. The backup parameters for the office scenario are 5000K / 45% / 400lux, the backup parameters for the medical scenario are 3500K / 30% / 120lux, and the backup parameters for the commercial scenario are 4000K / 60% / 500lux.

[0106] The following provides a detailed description of the LED color-emotion association map construction method and intelligent lighting control system of the present invention:

[0107] The LED color-emotion association map construction method of the present invention is based on "controlled variable experiments + multimodal data collection + machine learning modeling", and the specific steps are as follows:

[0108] Step S1: Construct a multi-dimensional emotion assessment system;

[0109] Core emotional dimensions: Pleasure (P), Arousal (A), Fit (S), Dominance (D), using the internationally recognized PAD emotional scale (1-7 points), where:

[0110] Pleasure level (P): 1 point = extreme annoyance, 7 points = extreme pleasure, quantified by the "Visual Analog Scale (VAS)" (0-100mm scale, user marks the corresponding position);

[0111] Arousal level (A): 1 point = extremely calm, 7 points = extremely excited, and the "emotional trigger delay time" (the time from the presentation of the color stimulus to the user's rating, accurate to 100ms) is recorded simultaneously.

[0112] Fit (S): 1 point = completely mismatched to the scene, 7 points = completely matched. The weight is set according to the functional requirements of the scene (such as "sleep aid" in the bedroom, "focus" in the office).

[0113] Advantage (D): 1 point = completely passive acceptance, 7 points = active control. For adjustable lighting products, a new indicator of "willingness to adjust parameters" has been added (1 point = unwilling to adjust, 7 points = active adjustment).

[0114] Derivative metrics: duration of emotion (the duration of the emotional state after the color stimulus stops, recorded to the minute level), and rate of emotion decay (the rate at which the emotion score drops from its peak to its baseline, in minutes / minute).

[0115] Step S2: Determine the full-dimensional color parameters and control precision of the LED (add hardware parameters);

[0116] The core parameters and technical requirements are shown in Table 1 below:

[0117]

[0118] Table 1

[0119] Interference parameter control: Fixed color rendering index Ra≥90 (R9≥80, to ensure color reproduction) and flicker depth≤1% (compliant with GB / T31168-2014 standard) to avoid non-target parameters affecting emotion assessment.

[0120] Step S3: Experimental design and data acquisition specifications for the controlled variable method;

[0121] Building a standardized testing environment:

[0122] Space requirements: The test chamber measures 3m × 3m × 2.8m, with the inner walls finished in matte neutral gray paint. To avoid glare;

[0123] Ambient light control: The system uses blackout curtains and flicker-free background lights to keep the ambient illuminance stable at ≤50 lux, the ambient color temperature is fixed at 5000K (D50 standard light source), the humidity is 40%-60%, and the temperature is 23±2℃.

[0124] Equipment deployment: The LED test light source is installed at the center of the top of the test chamber (2.5m above the ground). The subject's seat height is adjustable (ensuring a vertical distance of 60±5cm between the line of sight and the light source). The physiological signal acquisition device is placed to the side of the seat (without electromagnetic interference).

[0125] Subject screening and grouping:

[0126] Sample size: 300 people in total, divided into 6 groups (50 people each in the age groups of 18-25, 26-35, 36-45, 46-55, and 56-60, with an equal number of men and women).

[0127] Inclusion criteria: passing the Ishihara color vision test (no red-green color weakness / color blindness), no mood disorders such as anxiety / depression, and no use of medications affecting the nervous system in the past 3 months;

[0128] Cultural background coverage: Eastern culture group (100 people from China), Western culture group (80 people from the United States), Southeast Asian culture group (60 people from India), Middle Eastern culture group (60 people from the United Arab Emirates) to ensure the validity of cross-cultural adaptation data.

[0129] Experimental procedure specifications:

[0130] Pretreatment phase: After entering the test chamber, the subjects first undergo 10 minutes of environmental adaptation (viewing a neutral gray screen to avoid initial emotional fluctuations), and basic physiological signals (skin conductivity, heart rate) are collected simultaneously as a baseline;

[0131] Parameter presentation method: "Random sequence presentation" is adopted (random parameter combination sequence is generated by Matlab), each parameter combination is presented for 10 seconds (to ensure stable emotion perception), with an interval of 3 seconds (presenting a neutral gray screen to avoid emotion superposition);

[0132] Data acquisition synchronization: Synchronize the following data via timestamps:

[0133] 1. Subjective rating: After the parameters are presented, the subjects complete the four-dimensional rating of PAD through the touch screen (with a time limit of 10 seconds to avoid bias).

[0134] 2. Physiological signals: skin conductivity (sampling frequency 10Hz, electrode attached to the fingertip of the index finger), heart rate (chest patch heart rate belt, sampling frequency 5Hz), cortisol level (saliva samples were collected once before and once after the test, and detected by enzyme-linked immunosorbent assay).

[0135] 3. Neuroelectric signals: A 64-channel EEG cap (sampling frequency 250Hz) was used to collect signals from the prefrontal lobe (F3 / F4), temporal lobe (T3 / T4), and occipital lobe (O1 / O2) regions. The amplitude and latency of the P1 (80-120ms, early visual processing), N1 (150-200ms, emotional arousal), and LPP (300-600ms, sustained emotional processing) bands were analyzed.

[0136] Repeated test design: Each subject needs to complete 3 rounds of testing (each round is 24 hours apart). The third round of testing adds a "parameter memory" assessment (to determine whether the parameter combination has been seen before), and removes memory bias data (samples with memory ≥80%).

[0137] Step S4: Multimodal data processing and mapping relationship modeling;

[0138] Data preprocessing:

[0139] Outlier removal: using " "Criteria + manual verification" first removes those exceeding the mean. Abnormal data (such as a sudden spike in skin conductivity to over 10 μS, which may be due to interference from the subject's movements) are then manually checked for contradictory scoring data (such as samples with pleasure score of 7 but arousal score of 1).

[0140] Data standardization:

[0141] Physiological signals: Z-score normalized ( (Eliminating individual baseline differences, such as the influence of higher baseline heart rate in older adults).

[0142] EEG signal: Filtered using EEGLAB software (0.5-30Hz bandpass filtering), removal of electrooculography artifacts (ICA independent component analysis), and extraction of average amplitude for each band.

[0143] Data alignment: Using the start time of parameter presentation as the timestamp origin, subjective scores, physiological signals, and EEG signals are aligned along the time axis to ensure the temporal correlation between "parameter stimulus and emotional response".

[0144] Correlation analysis methods:

[0145] Correlation analysis: The Pearson correlation coefficient method was used to analyze the strength of the association between a single color parameter and the emotion dimension. For example:

[0146] Color temperature and wakefulness: r = -0.78 (P < 0.001), indicating that as the color temperature increases (warm light → cool light), the wakefulness decreases significantly;

[0147] Saturation and pleasure: r=0.65 (P<0.001), indicating that when the saturation is in the range of 0%-70%, the pleasure increases with the increase of saturation, and shows a downward trend after exceeding 70% (r=-0.32).

[0148] Multiple regression modeling: Construct a multiple linear regression model for "color parameter combination → emotion dimension", with the following formula:

[0149] Pleasantness (P) = a1 × color temperature + a2 × saturation + a3 × brightness + a4 × spectral peak value + b (constant term);

[0150] Where a1-a4 are regression coefficients (fitted using the least squares method; for example, a1=-0.002, meaning that for every 100K increase in color temperature, the pleasure score decreases by 0.2 points), and the model fit is... ;

[0151] Physiological indicator threshold calibration: For different emotional goals, determine physiological signal thresholds, for example:

[0152] "Relaxation and Sleep Aid": Skin conductivity ≤3μS, Heart rate ≤70 beats / minute Band (8-13Hz) amplitude ≥8μV;

[0153] "Focus on High Efficiency": Skin conductivity 3-5μS, heart rate 70-80 beats / minute, EEG β band (14-30Hz) amplitude ≥6μV.

[0154] Step S5: Construction and retrieval optimization of the three-dimensional association map;

[0155] Atlas structure design:

[0156] Coordinate axis definition: X-axis is "core color parameter combination" (coded by color temperature-saturation-brightness, such as 2700K_50%_60lux), Y-axis is "emotional dimension quantification value" (PAD four-dimensional score + derived indicators), Z-axis is "physiological signal threshold range" (skin conductivity-heart rate-EEG band amplitude).

[0157] Layer division: Divided into "Home Layer", "Office Layer", "Commercial Layer" and "Medical Layer" according to application scenarios. Each layer is labeled with the emotional target weight in that scenario (e.g., "comfort" weight is 0.6 and "pleasure" weight is 0.3 in the medical layer).

[0158] Confidence interval labeling: Each parameter-emotion mapping point is labeled with a 95% confidence interval (e.g., 2700K color temperature corresponds to a pleasure level of 6.2±0.3 points). Mapping points with a confidence level below 90% are labeled "to be verified" to avoid misleading applications.

[0159] Spectrum search function:

[0160] Multi-dimensional search: Supports "emotional goal search" (e.g., input "relaxation and sleep aid", output corresponding parameter combinations), "scene search" (e.g., input "bedroom", output parameter schemes for all emotional goals in that scene), and "physiological signal search" (e.g., input "heart rate 65 beats / minute", output matching emotion-parameter combinations);

[0161] Personalized correction: Supports inputting user characteristics (age, cultural background, color vision type) and automatically correcting parameter schemes. For example, if a 56-year-old East Asian user selects "relaxation and sleep aid", the system will automatically adjust the color temperature from 2700K to 2800K (to compensate for the reduced blue light perception of the elderly) and increase the saturation from 50% to 55%.

[0162] The intelligent lighting control system of this invention adopts a collaborative architecture of "hardware + software + algorithm". The functions and technical parameters of each module are as follows:

[0163] Data acquisition module;

[0164] Scene recognition unit:

[0165] Hardware components: 2-megapixel RGB camera (supports 1080P@30fps) + infrared human body sensor (detection distance 0.5-5m);

[0166] Recognition Algorithm: The YOLOv5 object detection + scene classification model is used (the training set contains 100,000 images of home / office / commercial scenes). The recognition categories include "Bedroom (bed / wardrobe)", "Office (workstation / conference table)", "Retail store (shelf / cashier)" and "Hospital ward (bed / medical equipment)". The recognition accuracy is ≥95% and the recognition latency is ≤300ms.

[0167] Calibration mechanism: The system automatically collects images of the current scene every 7 days and compares them with the initial scene model. If the matching degree is less than 85% (such as adding furniture to the bedroom), recalibration is triggered.

[0168] Ambient light detection unit:

[0169] Hardware selection: Spectral sensor (model: AMSTCS34725, detection wavelength 400-700nm) + Illuminance sensor (model: VEML7700, detection range 0-1000lux).

[0170] Detection accuracy: color temperature detection error ±100K, illuminance detection error ±5%, sampling frequency 1Hz, storing ambient light data once every 5 minutes to construct an "ambient light change curve".

[0171] Physiological signal acquisition unit:

[0172] Wearable device compatibility: Supports Bluetooth 5.0 connection to smart bracelets (collecting heart rate and skin conductivity) and smart headbands (collecting simplified EEG signals, 3 electrodes), with a sampling frequency of 10Hz.

[0173] Embedded data acquisition: For fixed scenarios (such as office workstations), an integrated patch-type electrodermal sensor (electrode material: Ag / AgCl, impedance) is used. Powered via USB-C interface, data transfer latency ≤100ms;

[0174] Abnormal data handling: When physiological signals are interrupted (such as when the bracelet disconnects), it automatically switches to "scene default parameters" and sends a notification via the APP.

[0175] Association map storage module;

[0176] Storage architecture: It adopts a dual storage system of "local + cloud". The local storage stores the basic map (about 50MB, which can be used offline), and the cloud storage stores the complete map (including multi-scenario iterative data, about 200MB).

[0177] Data security: Local data is encrypted using AES-256, cloud data complies with the ISO27001 information security standard, and user data is anonymized (only associated with device ID, not personal identification information).

[0178] Version Management: The map is maintained on an annual iteration + quarterly update basis. The version number format is "V year.quarter" (e.g., V2025.1). Historical versions are retained during updates, and rollback is supported (if users report that the parameters of the new version are not compatible, they can switch to the old version).

[0179] Emotional needs matching module;

[0180] The scene-emotion mapping library is shown in Table 2 below:

[0181]

[0182] Table 2

[0183] Multi-factor matching algorithm:

[0184] i. Basic matching: Based on the scene recognition results, extract the parameter combinations corresponding to the core emotion target from the graph;

[0185] ii. Individual correction: Input user characteristics (e.g., "50-year-old male / Eastern culture / normal color vision"), and call the correction coefficient library (e.g., color temperature +100K, saturation +5%).

[0186] iii. Environmental adaptation: Compare the current ambient light parameters with the standard environment (5000K / 50lux) and calculate the correction value (e.g., if the ambient color temperature is 6000K, the LED color temperature will be automatically reduced by 200K to maintain the mood effect).

[0187] iv. Conflict handling: When multiple emotional goals conflict (e.g., the need for "party enjoyment" and "children's relaxation" exists simultaneously in the living room), prioritize them (prioritize level 1 goals) or use "parameter fusion" (e.g., take the middle value of the two for saturation).

[0188] Lighting control module;

[0189] LED driver unit:

[0190] Hardware components: MCU (model: STM32F407, main frequency 168MHz) + four-channel LED driver chip (model: TITPS92661, supports constant current output 0-1A).

[0191] Control precision: Color temperature adjustment error ±50K, saturation adjustment error ±1%, brightness adjustment error ±5lux, response time ≤100ms;

[0192] Dimming method: A hybrid mode of "PWM + linear dimming" is adopted. Linear dimming is used for low brightness (<100 lux) (to avoid flicker), and PWM dimming is used for high brightness (to ensure efficiency).

[0193] Color Management Unit:

[0194] Color space conversion: Supported Real-time conversion with a conversion error ≤ΔE0.5, using 32-bit floating-point arithmetic to improve accuracy;

[0195] Spectral calibration: The spectral sensor is automatically invoked every 24 hours to detect the deviation between the actual output spectrum and the theoretical spectrum. If the deviation is >10%, the RGBW current ratio is automatically adjusted to ensure stable color parameters.

[0196] Multi-zone coordinated control:

[0197] Zone management: Supports up to 16 lighting zones (e.g., the living room is divided into a main lighting zone and an ambient lighting zone), with independent parameter control for each zone;

[0198] Linkage logic: Set the "main partition - slave partition" relationship (e.g., if the main light in the bedroom is set to "sleep aid mode", the ambient light in the slave partition will automatically match the same color scheme with low saturation parameters), and the linkage delay is ≤500ms.

[0199] Dynamic optimization module;

[0200] Feedback data collection:

[0201] Proactive feedback: Users can complete an "emotional satisfaction rating" (1-5 points) via the APP / touchscreen and can add text feedback (such as "too bright" or "color too cool").

[0202] Passive feedback: Indirect data is collected through sensors (e.g., if the user's heart rate is consistently >85 beats / minute in "focus mode", it is judged as "emotional mismatch").

[0203] Iterative Algorithm:

[0204] Short-term optimization: The "gradient descent method" is adopted, and the parameters are adjusted according to the single feedback (e.g., satisfaction score of 3 points, color temperature reduced by 100K, saturation ±5%). The iteration step size decreases as the number of feedbacks increases (to avoid parameter fluctuations).

[0205] Long-term optimization: Using the "random forest algorithm", we analyze the correlation between user feedback data and parameter combinations every week and update the personalized parameter model (e.g., if users consistently report that "3000K is too warm", we automatically adjust the color temperature benchmark of "relaxation" mode to 3100K).

[0206] Group optimization: All user feedback is aggregated in the cloud, and the general graph is updated every quarter to improve group adaptability (e.g., if it is found that 80% of users aged 25-35 prefer the office parameters of "4500K+60% saturation", this combination will be set as the default).

[0207] User interaction module;

[0208] Interaction method:

[0209] Voice interaction: Supports Mandarin / English / Cantonese recognition, voice commands include "switch sleep aid mode" and "increase brightness by 50%", recognition accuracy is ≥92%, and offline commands are supported (10 commonly used commands);

[0210] APP Interaction: Offers "Professional Mode" (manually adjust color temperature / saturation / brightness sliders) and "Easy Mode" (select emotional target and switch with one click), supports parameter scheme collection (up to 50 sets) and sharing (generate QR code).

[0211] Personalized solutions:

[0212] User profiles: Generate user profiles based on age, cultural background, and usage habits (e.g., "28-year-old female / likes highly saturated warm colors / frequently used in bedroom scenes"), and automatically recommend suitable solutions;

[0213] Scheduled tasks: Supports setting "scene timed trigger" (such as "automatically switch to sleep aid mode in the bedroom at 22:00 every night") and "mood curve timed" (such as "in sleep aid mode, the color temperature gradually changes from 3000K to 2700K and the brightness decreases from 80 lux to 50 lux within 1 hour").

[0214] Fault self-diagnosis and redundancy modules;

[0215] Fault detection:

[0216] Hardware failure: Real-time monitoring of LED driver current and voltage. If the current exceeds the 0-1A range or the voltage fluctuation is >±10%, it is determined as "driver failure"; monitoring of spectral sensor data. If the acquisition fails 5 times in a row, it is determined as "sensor failure".

[0217] Parameter deviation: Detects the deviation between the actual output parameters and the target parameters. If the color temperature deviation is >150K, the saturation deviation is >10%, and the brightness deviation is >20lux, and this continues for 10 seconds, it is judged as "parameter out of control".

[0218] Troubleshooting:

[0219] Level 1 fault (such as parameter loss): Automatically switch to "backup parameter scheme" (pre-stored locally, such as 2700K / 50% saturation / 100lux), and push fault notification via APP;

[0220] Level 2 faults (such as drive failures): Turn off the lighting in the faulty zone and turn on the backup lighting (e.g., if the main bedroom light fails, automatically turn on the bedside ambient light), and generate a fault code (e.g., "ERR-DRV-01") for easy maintenance.

[0221] Fault Logs: Nearly 100 fault records are stored locally (including time, type, and processing result), and can be exported to Excel for easy product iteration analysis.

[0222] The LED color-emotion association map construction method and intelligent lighting control system of the present invention have at least the following advantages:

[0223] 1. Improved precision in emotion regulation:

[0224] Comparative experiment: In a sample of 300 people, the LED system using the method of this invention achieved an emotion target matching accuracy of 89%, which is 43.5% higher than the traditional fixed parameter system (62%).

[0225] Physiological indicators verified that under the "sleep aid mode", users' average time to fall asleep was shortened by 18 minutes and the number of awakenings per night decreased by 2.1 times (PSQI sleep quality index dropped from 12.3 to 7.8).

[0226] 2. Individual fit optimization:

[0227] Age suitability: For users aged 56-60, the emotional satisfaction rate reached 87% after parameter adjustment, which is 33.8% higher than the unadjusted solution (65%).

[0228] Cultural compatibility: 90% of Eastern users are satisfied with the red-themed parameters, and 85% of Western users are satisfied with the red-themed parameters (compared to 72% and 60% for unadapted solutions, respectively).

[0229] 3. Enhanced dynamic response capability:

[0230] Environmental adaptation: When the ambient light increases from 50 lux (cloudy day) to 500 lux (sunny day), the system parameter adjustment response time is ≤500ms, and the emotional satisfaction rate is maintained above 85% (compared to 68% for traditional systems).

[0231] Emotional fluctuation adaptation: When the user's heart rate rises from 70 beats / minute to 90 beats / minute (anxiety), the system automatically switches to "soothing mode", and the heart rate drops by an average of 12 beats / minute after 5 minutes.

[0232] 4. Enhanced product application value:

[0233] In office settings: After adopting this system, users' work efficiency increased by 19% (task completion time was reduced by 22 minutes / day), and the Visual Ability Score (VAS) decreased from 6.2 to 3.1;

[0234] In the healthcare setting: After application in hospital wards, patients' anxiety scale (GAD-7) scores dropped from 8.5 to 5.2, and the number of nighttime calls by nurses decreased by 35%.

[0235] The following are specific embodiments of the LED color-emotion association map construction method and intelligent lighting control system of the present invention.

[0236] Example 1: Home bedroom scene (sleep aid mode).

[0237] I. System Deployment.

[0238] Lighting zones: Main light zone (ceiling light, RGBW four-color LED), ambient light zone (bedside light strip, RGB LED);

[0239] Data collection devices: smart bracelet (collects heart rate and skin conductivity), scene recognition camera (installed in the corner of the bedroom), ambient light sensor (installed on the windowsill).

[0240] II. Work Process.

[0241] 1. Scene Recognition: At 19:30, the user enters the bedroom, and the camera recognizes "bed + wardrobe", which is determined to be a "bedroom scene". Based on the time (evening), the core emotional goal is set to "relaxation and sleep aid" (priority level 1).

[0242] 2. Data Acquisition: The ambient light sensor detects the current ambient color temperature as 4500K and illuminance as 300 lux; the smart bracelet collects the user's heart rate as 75 beats / minute and skin conductivity as 4.2μS.

[0243] 3. Parameter matching:

[0244] Basic parameters: Extracted from the spectrum: "Relaxation and Sleep Aid" basic parameters: color temperature 2700K, saturation 50%, brightness 60 lux;

[0245] Individual correction: The user is a 48-year-old male (Eastern culture), and the correction coefficients are: color temperature +100K (2800K), saturation +5% (55%).

[0246] Environmental adaptation: Ambient illuminance 300 lux > standard 50 lux, brightness increased to 80 lux to avoid discomfort caused by excessively bright ambient light;

[0247] 1. Lighting control: Main light output 2800K / 55% / 80lux, ambient light output 2800K / 40% / 30lux (from zone parameters), dimming frequency 25kHz, flicker-free;

[0248] 2. Dynamic optimization: At 22:00, the user reported that the brightness was "a bit too bright" (satisfaction score of 3), so the system reduced the brightness to 60 lux; at 23:00, the system detected that the user's heart rate was still 72 beats / minute (higher than the sleep aid threshold of 70 beats / minute), so the system automatically reduced the color temperature to 2700K, and the heart rate dropped to 68 beats / minute after 30 minutes.

[0249] 3. Troubleshooting: At 2:00 AM the next day, the main light drive current was abnormal (exceeding 1A). The system determined that there was a "drive failure", automatically turned off the main light, and turned on the ambient light (backup parameters 2700K / 50% / 50lux). The APP pushed an "ERR-DRV-01" fault message.

[0250] Example 2: Office workstation scenario (focus mode).

[0251] I. System Deployment.

[0252] Lighting zones: Main workstation light (RGBWLED), desk lamp (RGBLED);

[0253] Data acquisition devices: embedded skin conductance sensor (attached to the workstation desktop), EEG smart headband (worn by the user), ambient light sensor (installed above the workstation).

[0254] II. Work Process.

[0255] 1. Scene Recognition: At 9:00, the user is sitting at their workstation. The camera recognizes "computer + files" and determines it as an "office scene". The core emotional goal is set as "focus and efficiency".

[0256] 2. Data Acquisition: Ambient light color temperature 5500K, illuminance 400 lux; skin conductivity 3.8 μS, EEG β band amplitude 5.2 μV (below the focus threshold of 6 μV);

[0257] 3. Parameter matching:

[0258] Basic parameters: color temperature 5000K, saturation 45%, brightness 400 lux;

[0259] Individual adjustment: The user is a 28-year-old female (Western culture), with a red-themed adjustment coefficient of 0.9, and other parameters remain unchanged;

[0260] Neural electrical correction: Insufficient amplitude of EEG β band, brightness increased to 450 lux, promoting β band activation;

[0261] 1. Lighting control: Main lamp output 5000K / 45% / 450lux, desk lamp output 5000K / 50% / 300lux (enhancing desktop illuminance);

[0262] 2. Dynamic optimization: Detected at 14:30 When the band amplitude increases to 9μV (attention decreases), the system raises the color temperature to 5500K and reduces the saturation to 40%. After 10 minutes, the β band amplitude increases to 6.5μV. At 16:00, the user reports "eye fatigue" (satisfaction score 2 points), so the brightness is automatically reduced to 400 lux and "low-frequency flicker mode" (dimming frequency 50kHz) is activated.

[0263] Example 3: Medical ward scenario (comforting mode).

[0264] I. System Deployment.

[0265] Lighting zones: Ward main light (high CRI RGBWLED, Ra≥95), bedside light (dimmable RGBLED);

[0266] Data acquisition equipment: medical-grade heart rate monitor (connected to patient monitoring system), ambient light sensor (antibacterial housing).

[0267] II. Work Process.

[0268] 1. Scene Recognition: After a patient is admitted to the hospital post-surgery, the system obtains "post-operative recovery period" information through the hospital's HIS system and determines that the core emotional goal is "soothing and calming".

[0269] 2. Data acquisition: Ambient light color temperature 4000K, illuminance 200 lux; patient heart rate 85 beats / minute (above the resting threshold of 75 beats / minute);

[0270] 3. Parameter matching:

[0271] Basic parameters: color temperature 3500K, saturation 30%, brightness 150 lux (low saturation to avoid visual stimulation);

[0272] Medical correction: Post-operative patients are sensitive to strong light, so the brightness is reduced to 120 lux and the spectral peak is adjusted to 550nm (green light band, which has a calming effect).

[0273] 1. Lighting control: Main light output 3500K / 30% / 120lux, bedside light output 3500K / 25% / 80lux;

[0274] 2. Dynamic optimization: After 2 hours, the patient's heart rate dropped to 78 beats / minute, and the system maintained the parameters; after 4 hours, the patient reported "wanting to read" (satisfaction score of 3), and the bedside lamp brightness was automatically increased to 200 lux, saturation increased to 35%, and color temperature was maintained at 3500K (to avoid sudden color temperature stimulation).

[0275] Supplementary explanation of the specific implementation method:

[0276] 1. Parameter expansion: The color parameters in the embodiments of the present invention can be expanded to "pulse frequency" (such as 20Hz-100Hz pulsed light, used for pain relief in medical scenarios) and "spectral continuity" (such as full-spectrum LED, to improve emotional comfort).

[0277] 2. Cross-device adaptation: Supports integration with smart home systems (such as Xiaomi Mijia and Huawei HarmonyOS) to achieve multi-device emotional scene coordination for "lighting + curtains + music";

[0278] 3. Special population adaptation: For users with color vision deficiency, a "color vision deficiency compensation mode" can be added (such as increasing the proportion of yellow band for users with red-green color vision deficiency), and for visually impaired users, a "voice parameter broadcast" function can be added.

[0279] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for constructing an LED color-emotion association map, characterized in that, Includes the following steps: S1: Construct a multi-dimensional emotion assessment system, which includes core emotion dimensions and derived indicators. The core emotion dimensions include pleasure, arousal, fit, and dominance. The derived indicators include emotion duration and emotion triggering delay time. The core emotional dimensions are quantified using a 1-7 point PAD emotional scale. The pleasantness is quantified with the help of the Visual Analogue Scale (VAS). The arousal level records the emotional trigger delay time in sync. The suitability is weighted in combination with the functional requirements of the scene. The advantage level is related to the parameter adjustment of the willingness index. S2: Determine the full-dimensional color parameters of the LED light source, including color temperature, saturation, brightness, color gamut type, color deviation value, dynamic gradient speed, and spectral peak wavelength; The color temperature range is 2200K-6500K, with an adjustment accuracy of ±50K; The saturation range is 0%-100%, with an adjustment accuracy of ±1%. The brightness is divided into lighting scenarios (50 lux-800 lux, accuracy ±5 lux) and display scenarios (50 nit-500 nit, accuracy ±2 nit). The color gamut types include sRGB, DCI-P3, and AdobeRGB, with a coverage of ≥98%; The color deviation values ​​include general-grade DeltaE≤3 and professional-grade DeltaE≤1, with an adjustment accuracy of ±0.

2. The range of the dynamic gradual speed change is 0.5s-5s, with an accuracy of ±0.1s; The peak wavelength of the spectrum is in the range of 400nm-760nm, with an accuracy of ±5nm; S3: Design an experiment using the controlled variable method and collect multimodal data. The experiment should be conducted in a standardized testing environment. The standardized testing environment meets the following requirements: spatial dimensions of 3m × 3m × 2.8m, with matte neutral gray inner walls. Ambient illuminance ≤50 lux, D50 standard light source color temperature 5000K, humidity 40%-60%, temperature 23±2℃; A total of 300 participants were divided into 6 age groups, with 50 participants aged 18-25 and 50 participants aged 56-60, half male and half female, covering 4 cultural backgrounds: East, West, Southeast Asia and Middle East, and all of them passed the Ishihara color vision test. The experiment uses a random sequence to present color parameter combinations, ensuring that each combination is presented for 10 seconds with a 3-second interval, and that data is collected synchronously. Subjective rating, which is a 10-second timed rating based on the four dimensions of PAD; Physiological signals, namely skin conductivity sampled at 10 Hz, heart rate sampled at 5 Hz, and salivary cortisol sampled once before and once after the test; The neural electrical signals were sampled at 250Hz using a 64-channel EEG cap, with a focus on acquiring the F3 / F4, T3 / T4, and O1 / O2 regions. Each participant completed 3 rounds of testing, with each round spaced 24 hours apart. The third round included an additional parameter: memory assessment. S4: Multimodal data processing and establishing mapping relationships, including: Outlier removal, the outlier removal is as follows: The criteria and the logic of manual verification contradict each other; Data standardization, which includes Z-score processing of physiological signals, 0.5-30Hz bandpass filtering of EEG signals, and ICA artifact removal; Data alignment, wherein the data alignment synchronizes three types of data according to timestamps; The Pearson correlation coefficient method was used to analyze the association between parameters and emotions. A strong association was defined as r ≥ 0.7, P < 0.

001. A multiple linear regression model was constructed, and physiological index thresholds were calibrated. The goodness of fit of the multiple linear regression model was then analyzed. The physiological indicator thresholds include relaxation and sleep aid, corresponding to skin conductivity ≤3μS, heart rate ≤70 beats / minute, and EEG α band amplitude ≥8μV; S5: Construct a three-dimensional correlation map. The three-dimensional correlation map uses the core color parameter combination composed of color temperature, saturation, and brightness as the X-axis, the emotional dimension quantification composed of pleasure, arousal, fit, and dominance as the Y-axis, and the physiological signal threshold range composed of physiological signals as the Z-axis. It is divided into home layer, office layer, commercial layer, and medical layer according to the scene. Each mapping point is marked with a 95% confidence interval. The confidence level <90% is marked "to be verified". The map supports multi-dimensional retrieval of emotional goals, scenes, and physiological signals, and can automatically correct parameters by inputting user characteristics.

2. The method for constructing an LED color-emotion association map according to claim 1, characterized in that, In step S3, the subject needs to undergo 10 minutes of environmental adaptation by watching a neutral gray screen before the test, and basic physiological signals are collected simultaneously as a baseline; during EEG signal analysis, the amplitude and latency data of the P1 band, N1 band, and LPP band are extracted.

3. The method for constructing an LED color-emotion association map according to claim 1, characterized in that, In step S4, the formula for the multiple linear regression model is: Pleasure (P) = a1 × color temperature + a2 × saturation + a3 × brightness + a4 × spectral peak value + b, where a1-a4 are regression coefficients and b is a constant term.

4. An intelligent lighting control system, characterized in that, The three-dimensional association map obtained by applying the LED color-emotion association map construction method according to claims 1-3 includes: The data acquisition module includes a scene recognition unit, an ambient light detection unit, and a physiological signal acquisition unit; The scene recognition unit uses a 2-megapixel RGB camera and an infrared sensor to recognize six types of scenes, including bedrooms and offices, through YOLOv5 and a scene classification model, and is automatically calibrated every 7 days. The ambient light detection unit uses an AMSTCS34725 spectral sensor and a VEML7700 illuminance sensor, with detection errors of color temperature ±100K and illuminance ±5%, sampling at 1Hz and constructing an ambient light change curve. The physiological signal acquisition unit supports Bluetooth 5.0 connection to wearable devices or integrates Ag / AgCl patch sensors, and automatically switches to the default scene parameters when the signal is interrupted. The associated map storage module includes dual storage: local and cloud storage. The local storage contains 50MB of basic map data, while the cloud storage contains 200MB of complete map data. Local data is encrypted with AES-256, and cloud data complies with the ISO27001 standard, ensuring user data anonymization. The emotion needs matching module has a built-in scene-emotion mapping library and uses a multi-factor matching algorithm: basic matching, individual correction, environmental adaptation, and conflict handling. The lighting control module includes an LED driver unit and a color management unit; The LED driving unit uses an STM32F407MCU and a TITPS92661 four-channel driving chip, with constant current 0-1A output, adjustment error of color temperature ±50K, saturation ±1%, brightness ±5lux, response ≤100ms, and adopts PWM and linear hybrid dimming. The color management unit supports Conversion, spectral calibration every 24 hours; supports 16-zone linkage; The dynamic optimization module collects both active and passive feedback; it uses gradient descent for short-term optimization and random forest algorithm for long-term optimization, with the population iterating over a general graph in the cloud every quarter. The user interaction module supports voice and APP interaction, and is used to generate user profiles and recommend solutions. It supports 50 sets of solutions to be collected and shared, or setting a timer for the emotion curve. The fault self-diagnosis module is used to monitor the drive current, voltage, and spectrum sensors in real time to determine hardware faults, or to monitor parameter deviations to determine parameter out of control; for first-level faults, it switches to backup parameters and provides a prompt; for second-level faults, it shuts down the fault zone, starts backup lighting, and generates a fault code; 100 fault records can be stored locally and exported.

5. The intelligent lighting control system according to claim 4, characterized in that, The scenario-emotion mapping library includes the emotional goals and weight settings for retail store windows, hospital wards, and living room gatherings.

6. The intelligent lighting control system according to claim 4, characterized in that, The emotional needs matching module includes a cultural difference adaptation unit. In the cultural difference adaptation unit, the correction coefficient for the red series parameters is 1.1-1.2 for Eastern culture, 0.8-0.9 for Western culture, 0.95-1.05 for Southeast Asian culture, and 1.0-1.1 for Middle Eastern culture.

7. The intelligent lighting control system according to claim 4, characterized in that, The lighting control module can support multi-area linkage control with two trigger modes, including timed triggering and scene linkage triggering.

8. The intelligent lighting control system according to claim 4, characterized in that, The iteration step size of the dynamic optimization module decreases as the number of feedbacks increases. When the number of feedbacks for the same parameter combination is ≥10 and the satisfaction score is ≥4.5, the parameter combination is set as the default parameter for the corresponding scenario.

9. The intelligent lighting control system according to claim 4, characterized in that, The user interaction module's APP supports a family sharing function, which can add at least 5 family member accounts and is used to store personalized plans and support automatic switching of corresponding plans according to members' schedules.

10. The intelligent lighting control system according to claim 4, characterized in that, The backup parameter schemes of the fault self-test module are pre-stored according to scenario categories. The backup parameters for office scenarios are 5000K / 45% / 400lux, the backup parameters for medical scenarios are 3500K / 30% / 120lux, and the backup parameters for commercial scenarios are 4000K / 60% / 500lux.

Images