Illumination control method based on artificial intelligence

Abstract

The invention relates to the technical field of intelligent lighting control, and discloses a lighting control method based on artificial intelligence. Pre-defined driving current is applied to warm white and cold white channels of the two-color LED respectively, and color temperature is measured, so that accurate linear mapping is established; reasonably setting and checking sunshine starting and ending moments, color temperature upper and lower limits, a sampling interval, a moving average window and a total driving current, generating a sampling moment sequence in an effective sunshine period, and calculating a dynamic color temperature target according to a circadian rhythm; defining a light mixing function based on a color temperature-current model and a target curve, accurately solving a current distribution proportion through a mathematical method, and smoothing the current through moving average; and outputting a final current instruction at each sampling moment, automatically powering off at the end of the day, and restarting the next day.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent lighting control, and in particular to a lighting control method based on artificial intelligence. Background Art

[0002] In modern buildings and interior environments, lighting systems must not only fulfill basic illumination functions but also increasingly balance energy conservation, high efficiency, and human health. Traditional lighting control relies primarily on fixed-time on / off switching or manual adjustment. Color temperature and brightness parameters are often set based on experience, lacking intelligent response to circadian rhythms and individual needs. If the preset time periods are out of sync with actual usage scenarios, the lighting environment often mismatches human circadian rhythms, affecting comfort and work efficiency. Furthermore, using a single color temperature or single-channel control makes it difficult to achieve a more natural and healthy spectral adjustment.

[0003] Many existing smart lighting solutions rely on predefined schedules or simple ambient light sensors for automatic switching, such as automatically switching to high brightness during rush hour and dimming during idle periods. While these approaches offer some energy savings, they lack precise control over dynamic color temperature changes and are unable to adjust in real time to align with human circadian rhythms. Other systems simply operate warm and cool white lights in parallel, varying the color temperature through a fixed ratio or manual switching. However, these systems lack precise calibration and mathematical modeling for each channel, resulting in significant deviations from expected color temperature output. In the field of dual-color LED mixing, the industry's common practice is to assign color temperature based on empirical formulas or fixed dimming tables in the vehicle's DALI protocol. These solutions struggle to ensure smooth and synchronized changes in color temperature and brightness in complex scenarios, and lack online optimization based on real-time environmental conditions or user needs. Furthermore, most existing color temperature calibration methods only perform linear fitting on a single channel, ignoring the potential non-ideal characteristics between LED drive current and color temperature. Improving color temperature accuracy requires multi-point measurement and segmented fitting, which increases cost and engineering effort in actual deployment. In recent years, "human-centric lighting" technology has garnered widespread attention. Using biological rhythmic sinusoidal curves to simulate the temporal variations in natural light color temperature, it has been widely adopted in smart homes and commercial lighting. However, most commercial systems rely on preset curves and remote app control, lacking autonomous calculations and real-time adjustment capabilities. They still require manual design of time periods and spectral mapping, and lack intelligent response to specific environmental changes. Furthermore, these systems initiate closed-loop control upon startup, resulting in high operational complexity and stringent hardware performance requirements, making them unsuitable for resource-constrained embedded platforms. Regarding filtering and smoothing, existing control systems typically employ simple low-pass filters or fixed-parameter PID algorithms for brightness adjustment to reduce flicker and sudden changes. However, because model parameters are often empirically derived and fail to fully account for insufficient samples during system startup and at the end of the system, filter response lag or initial flicker often occur, impacting the user experience. Furthermore, the output of multi-channel light mixing often ignores the validity of real-time judgment results. If a calculation anomaly occurs, the system requires manual intervention or simply persists to the previous state, failing to adaptively degrade to a usable state.

[0004] To this end, this case aims to propose an AI-based lighting control method. Through a series of intelligent algorithms such as system modeling, parameter verification, target color temperature generation, proportional distribution, and dynamic output, the lighting equipment can automatically adjust the output color temperature according to circadian rhythms, environment, and usage needs, thereby improving the intelligence level of lighting and human comfort. Summary of the Invention

[0005] The present invention provides a lighting control method based on artificial intelligence, which helps solve the problems mentioned in the above background technology.

[0006] The present invention provides the following technical solution: a lighting control method based on artificial intelligence, comprising:

[0007] S1. Apply predefined driving currents to the warm white channel and the cool white channel of the dual-color LED, respectively, and establish a linear calibration model between the color temperature of each channel and the driving current based on the measured output color temperature;

[0008] S2. Define and verify system parameters, including sunshine start time, sunshine end time, target color temperature upper and lower limits, sampling time interval, sliding average filter window length, and total driving current of the two-color channel. Also, perform rationality check on the effective sunshine duration and determine the number of sampling times.

[0009] S3. Generate a sampling time sequence based on the effective sunshine duration and sampling interval;

[0010] S4. Based on the sampling time, according to the preset diurnal variation pattern, the target color temperature curve at each moment is calculated to obtain the dynamic color temperature target sequence of the lighting output within the full cycle;

[0011] S5. Based on the color temperature-current model of the two-color channel and the target color temperature curve, define a light mixing function, and solve the distribution ratio of the two-color channel current through a mathematical method;

[0012] S6. Allocate the current values ​​of each channel obtained and smooth the current data using a sliding average method;

[0013] S7. At each sampling moment, output the finally determined warm white and cool white channel driving current values ​​to the lighting driver;

[0014] S8. After the daily lighting cycle ends, the lighting power supply is automatically cut off and restarted the next day according to the established process.

[0015] Optionally, applying a predefined driving current to the warm white channel and the cool white channel of the dual-color LED, respectively, and establishing a linear calibration model between the color temperature of each channel and the driving current according to the measured output color temperature, specifically includes:

[0016] Two different drive currents are applied to the warm white LED channel and the cool white LED channel respectively:

[0017] Warm white channel current: I w1 , I w2 ; Corresponding measured color temperature: T w1 , T w2 ; Among them, I w1 I is the driving current of the first calibration point of the warm white channel; w2 is the driving current of the second calibration point of the warm white channel; T w1 For the warm white channel, the current I w1 Color temperature measured under Tw2 For the warm white channel, the current I w2 Color temperature measured under

[0018] Cold white channel current: I c1 , I c2 ; Corresponding measured color temperature: T c1 , T c2 ; Among them, I c1 I is the driving current of the first calibration point of the cold white channel; c2 is the driving current of the second calibration point of the cold white channel; T c1 For the cold white channel, the current I c1 Color temperature measured under T c2 For the cold white channel, the current I c2 Color temperature measured under

[0019] Calculate the gain α of the warm white channel color temperature change on the current w and the linear bias term b of the warm white channel color temperature and current w : Calculate the color temperature change corresponding to the unit current change of the warm white channel, reflecting the sensitivity of the drive current adjustment to the color temperature; b. w =T w1 -α w I w1 ;

[0020] Calculate the gain α of the cold white channel color temperature change on the current c and the linear bias term b of the cool white channel color temperature and current c : b c =T c1 -α c I c1 ;

[0021] Construct arbitrary warm white channel current I w Output color temperature function T under w (I w ) and any cold white channel current I c Output color temperature function T under c (I c ):

[0022] T w (I w )=α w I w +b w , T c (i c )=α c I c +b c .

[0023] Optionally, the definition and verification of system parameters, the rationality check of the effective sunshine duration and the determination of the number of sampling moments specifically include:

[0024] Preset by the designer: Sunshine start time t rise , end time of sunshine t set 、Target color temperature lower limit T min , target color temperature upper limit T max , sampling time interval Δt, sliding average filter window length M and the total driving current I fixedly allocated to the dual-color channel tot ;

[0025] Calculate the effective duration of sunshine D = t set -t rise If D≤0, it will prompt invalid sunshine duration, please reset it;

[0026] Calculate the number of sampling moments in a day

[0027] Optionally, generating a sampling time sequence based on the effective sunshine duration and the sampling interval specifically includes:

[0028] For each sampling point n = {0, 1, ..., N-1} calculate t[n] = t rise +nΔt; where n is the sampling number; t[n] is the nth sampling time;

[0029] If t[N-1]>t set , then let t[N-1]=t set .

[0030] Optionally, the step of calculating the target color temperature curve at each moment based on the sampling moment and in accordance with a preset diurnal variation pattern to obtain a dynamic color temperature target sequence of the lighting output within the full cycle specifically includes:

[0031] Calculate the normalized day-night phase at the nth moment

[0032] Get the target output color temperature T at the nth moment tgt [n]:

[0033] T tgt [n] = T min +(T max -T min )sin(φ[n]).

[0034] Optionally, defining a light mixing function based on a color temperature-current model of the dual-color channel and a target color temperature curve, and solving a distribution ratio of the dual-color channel current by a mathematical method, specifically includes:

[0035] Assuming the proportion of warm white channel current is r, the warm white channel current I w =r·I tot , cold white channel current I c =(1-r)·I tot ;

[0036] Construct mixed light color temperature function: Among them, T mix (r) is the warm and cool mixed output color temperature;

[0037] Expand and sort to get Ar 2 +Br+C=0;wherein, A=I tot (α w +α c ) is the coefficient of the quadratic term; B = b w -b c -2α c I tot is the coefficient of the first-order term; C=α c I tot +b c -T tgt [n] is a constant term;

[0038] Get the discriminant Δ=B 2 -4AC;

[0039] If Δ≥0, we get Among them, r1 and r2 are two roots of the quadratic equation, representing the possible solutions of the current proportion of the warm white channel;

[0040] Select the first root that satisfies 0≤r≤1 in turn. If both roots are there, take r1;

[0041] If Δ<0 or both roots are not in [0,1], it degenerates into a linear approximation:

[0042] in, This is the color temperature of the warm white channel alone at full current; is the color temperature of the cold white channel at full current; r[n] is the current ratio of the warm white channel at the nth sampling point;

[0043] like Then let r[n]=0.5;

[0044] Bounds cutoff:

[0045] Optionally, the obtained current values ​​of each channel are distributed and the current data are smoothed using a sliding average method, specifically including:

[0046] Original current distribution: I w [n]=r[n]Itot , I c [n]=(1-r[n])I tot ; Among them, I w [n] is the warm white channel current at the nth moment; I c [n] is the cold white channel current at the nth moment;

[0047] Get the number of valid samples K participating in the average at the nth moment n =min(M,n+1);

[0048] Calculate the smoothed current of the warm white channel at moment n

[0049] Calculate the smoothed current of the cold white channel at moment n

[0050] Optionally, at each sampling moment, outputting the finally determined warm white and cool white channel driving current values ​​to the lighting driver specifically includes:

[0051] At each sampling time t[n] write To LED driver;

[0052] If writing fails, the error log is recorded and the last valid value is maintained.

[0053] Optionally, the lighting power supply is automatically cut off after the daily lighting cycle ends, and restarted the next day according to a predetermined process, specifically including:

[0054] If the last sampling point t[N-1]≥t set , then write (0,0) to turn off the power of the lamp;

[0055] The system enters low power sleep mode until the next day rise Automatically returns to step S3 and starts again.

[0056] The present invention has the following beneficial effects:

[0057] 1. A precise linear model between drive current and output color temperature was systematically established for each channel of the bi-color LED. By applying different currents to the warm-white and cool-white channels and measuring the output color temperature, precise quantification of physical parameters was achieved. Unlike traditional lighting control methods that often rely on empirical values ​​or simple lookup tables, this method provides a highly reproducible mathematical foundation for subsequent light mixing algorithms. This improves the predictability and accuracy of lighting color temperature adjustment, avoiding problems such as color temperature drift and unstable output. It also addresses the issue of inaccurate adjustment caused by factors such as LED batch differences and temperature drift, laying a solid foundation for the system's adaptability and large-scale deployment.

[0058] 2. Incorporating the definition of system parameters, boundary verification, and anomaly warnings into the overall algorithm process ensures parameter compliance and operational safety of the lighting control system. Compared with the traditional practice of manual parameter adjustment by maintenance or design personnel and the lack of an automatic verification mechanism, this invention automatically calculates the effective sunshine duration and the number of sampling points, and provides real-time feedback for parameter setting anomalies (such as time overlap or unreasonable ranges), thereby improving the system's intelligent self-checking capabilities and configuration fault tolerance. This ensures the logical validity and boundary compliance of all subsequent adjustments and output processes, effectively preventing hidden dangers such as lighting anomalies, excessive energy consumption, or reduced comfort caused by parameter errors.

[0059] 3. An automated sampling time sequence generation mechanism has been introduced, ensuring that each sampling point for lighting adjustment is precisely aligned with the actual daylight rhythm. Unlike existing lighting systems, which only adjust at a few fixed time points or manually configure sampling points, this solution automatically assigns sampling time points throughout the day based on preset parameters and automatically corrects boundary points (such as the end of the day), effectively avoiding sudden changes in color temperature transitions and incorrect adjustments during invalid time periods. This improves the continuity of lighting adjustment and user experience, enabling the lighting system to seamlessly integrate with the human body's biological clock and natural light cycle, creating a more comfortable and healthy light environment.

[0060] 4. Based on the diurnal cycle and target color temperature range, a target color temperature curve is dynamically generated for each sampling moment, enabling continuous change of the lighting system over time. Unlike traditional solutions that only support segmented color temperature settings or simple day / night switching, this method accurately calculates the optimal color temperature for each moment of the day based on scientific human-centered lighting theory, improving human comfort and circadian synchronization. This not only meets the needs of natural light simulation and energy conservation in different application scenarios, but also automatically optimizes lighting output based on the human body's biological response to the light environment, enhancing the healthiness and intelligence of the lighting system.

[0061] 5. Utilizing physical modeling of the dual-color LED channels, the driving current distribution ratio between the two channels is accurately inferred for any target color temperature. The algorithm is also fault-tolerant and adaptable to extreme or abnormal scenarios. Unlike traditional dimming methods that rely on simple table lookups or manual interpolation, this method ensures that the output color temperature precisely falls within the target range at every moment through rigorous mathematical modeling and root cause identification. This improves the accuracy, response speed, and physical consistency of color temperature adjustment, avoiding common phenomena such as channel drive mismatch, brightness loss, and color temperature drift, and enhancing the scalability and adaptability of the lighting system in multiple scenarios.

[0062] 6. A sliding average algorithm is introduced into the dual-color channel current distribution, innovatively achieving a smooth transition in lighting color temperature output. Compared to previous lighting control systems, which are prone to flickering and sudden changes during channel switching or rapid response, this method effectively eliminates jitter and noise during adjustment through smoothing, improving visual comfort and lighting quality. This reduces the negative impact of sudden current changes on LED lifespan and user experience, making the lighting system more suitable for demanding environments (such as healthcare, education, and offices), while also facilitating the subsequent expansion of more complex dynamic lighting scenarios.

[0063] 7. Through automated communication with the driver, the calculated smoothed current value is regularly output to the actual hardware, and a closed-loop mechanism automatically records and handles output anomalies. Unlike traditional solutions, which can easily cause the entire system to go black or become unresponsive when an output failure occurs, this method automatically maintains the last valid value and logs it when a hardware communication failure occurs, effectively improving the robustness and reliability of the system. This reduces the risk of service interruption due to single points of failure, ensures long-term, stable, and high-quality system operation, and facilitates subsequent maintenance and fault tracking.

[0064] 8. This system achieves automatic closed-loop and energy-saving management for the lighting system. Unlike existing solutions that require manual power-off or timer-based power-off, this system automatically shuts down and enters low-power sleep mode after each daily lighting task, then automatically restarts the next day, without requiring manual intervention. This reduces system energy consumption and maintenance costs, extending equipment life and operational safety, making it particularly suitable for applications requiring long-term continuous operation and demanding energy efficiency. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 It is a schematic diagram of the process of the present invention. DETAILED DESCRIPTION

[0066] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0067] In one embodiment, a lighting control method based on artificial intelligence includes:

[0068] S1. Apply predefined driving currents to the warm white channel and the cool white channel of the dual-color LED, respectively, and establish a linear calibration model between the color temperature of each channel and the driving current based on the measured output color temperature;

[0069] S2. Define and verify system parameters, including sunshine start time, sunshine end time, target color temperature upper and lower limits, sampling time interval, sliding average filter window length, and total driving current of the two-color channel. Also, perform rationality check on the effective sunshine duration and determine the number of sampling times.

[0070] S3. Generate a sampling time sequence based on the effective sunshine duration and sampling interval;

[0071] S4. Based on the sampling time, according to the preset diurnal variation pattern, the target color temperature curve at each moment is calculated to obtain the dynamic color temperature target sequence of the lighting output within the full cycle;

[0072] S5. Based on the color temperature-current model of the two-color channel and the target color temperature curve, define a light mixing function, and solve the distribution ratio of the two-color channel current through a mathematical method;

[0073] S6. Allocate the current values ​​of each channel obtained and smooth the current data using a sliding average method;

[0074] S7. At each sampling moment, output the finally determined warm white and cool white channel driving current values ​​to the lighting driver;

[0075] S8. After the daily lighting cycle ends, the lighting power supply is automatically cut off and restarted the next day according to the established process.

[0076] By first applying predefined drive currents to the warm-white and cool-white channels of a dual-color LED and measuring the output color temperature, an accurate color temperature-current model is established, addressing the issue of inaccurate color temperature adjustment. Secondly, by rationally setting and verifying system operating parameters (such as sunlight start and end times, color temperature upper and lower limits, and sampling intervals), lighting anomalies and energy waste caused by improper parameter settings are effectively eliminated. Subsequently, sampling time sequence generation and target color temperature curve calculation ensure continuous and scientific color temperature adjustment, aligning with circadian rhythms and achieving comfortable, human-centric lighting. Furthermore, by accurately determining the mixing ratio and performing a sliding average of the channel currents, abrupt changes in lighting output and flicker are overcome, improving lighting quality. Finally, a timed output and automatic power-off restart mechanism ensure stable operation and energy efficiency. Overall, this approach not only enhances the intelligence of the lighting system, but also improves energy efficiency, user comfort, and system stability, making it ideal for offices, medical facilities, and educational institutions with demanding lighting requirements.

[0077] The method of applying a predefined driving current to the warm white channel and the cool white channel of the dual-color LED and establishing a linear calibration model between the color temperature of each channel and the driving current according to the measured output color temperature specifically includes:

[0078] Obtain the color temperature-current relationship of warm white and cool white channels to ensure accurate subsequent light distribution;

[0079] Two different drive currents are applied to the warm white LED channel and the cool white LED channel respectively:

[0080] Warm white channel current: I w1 , which is the data starting point for establishing the color temperature-current relationship for the warm white channel; I w2 , with I w1 Paired to obtain linear gain of the warm white channel; corresponding measured color temperature: T w1 , T w2 ; Among them, I w1 I is the driving current of the first calibration point of the warm white channel; w2 is the driving current of the second calibration point of the warm white channel; T w1 For the warm white channel, the current I w1 Color temperature measured under T w2 For the warm white channel, the current I w2 Color temperature measured under

[0081] Cold white channel current: I c1 , I c2 ; Corresponding measured color temperature: T c1 , T c2 ; Among them, I c1 I is the driving current of the first calibration point of the cold white channel; c2 is the driving current of the second calibration point of the cold white channel; respectively, they serve as the boundary data points of the color temperature-current relationship of the cold white channel; T c1 For the cold white channel, the current I c1 Color temperature measured under T c2 For the cold white channel, the current I c2 Color temperature measured under

[0082] T w1 , T w2 , T c1 , T c2 Measure the output color temperature of LED channels at different currents for calibration;

[0083] Calculate the gain α of the warm white channel color temperature change on the current w and the linear bias term b of the warm white channel color temperature and current w : Calculate the color temperature change corresponding to the unit current change of the warm white channel, reflecting the sensitivity of the drive current adjustment to the color temperature; b. w =T w1 -α w I w1 , determine the zero offset of the linear relationship between warm white channel color temperature and current to ensure that the function can accurately describe all intervals;

[0084] Calculate the gain α of the cold white channel color temperature change on the current c and the linear bias term b of the cool white channel color temperature and current c : b c =T c1 -α c I c1 ; The color temperature adjustment capability of the cool white channel can be reflected in an accurate mathematical relationship;

[0085] Construct arbitrary warm white channel current I w Output color temperature function T under w (I w ) and any cold white channel current I c Output color temperature function T under c (I c ):

[0086] T w (I w )=α w I w +b w , according to the input warm white current I w Calculate the actual color temperature in real time to provide accurate parameters for subsequent light mixing; c (I c )=α c I c +b c , corresponding to the cool white channel, ensuring that the color temperature of the two channels is linearly controllable with the current.

[0087] Detailed linear calibration of the LED dual-color channels solves problems in traditional lighting systems, such as inaccurate color temperature adjustment and output drift caused by LED characteristic fluctuations or batch differences. This enables the control system to dynamically allocate current based on the actual physical characteristics of each LED channel, achieving more precise color temperature control. Specifically, only by accurately understanding the response characteristics of the warm white and cool white channels can the current allocated to each channel at each moment in the light mixing process be accurately inferred, ultimately ensuring that the output color temperature always meets the target requirement. This not only improves the controllability of the light mixing effect but also provides a critical guarantee for the consistency and stability of lighting equipment in large-scale venues. Furthermore, this calibration mechanism greatly simplifies the complexity of subsequent system maintenance and upgrades, helping to enhance product competitiveness.

[0088] The definition and verification of system parameters, the rationality check of the effective sunshine duration and the determination of the number of sampling moments specifically include:

[0089] Clarify control parameters and ensure the legality of the operating period;

[0090] Preset by the designer: Sunshine start time t rise , end time of sunshine t set , set the effective time interval of daily lighting color temperature change, provide boundaries and target color temperature lower limit T for subsequent dynamic color temperature adjustment min , target color temperature upper limit T max , define the adjustment range of lighting color temperature so that the output color temperature is consistent with human rhythm and avoids crossing the limit, sampling time interval Δt, specify the system refresh and adjustment frequency, balance the response speed and system energy consumption, sliding average filter window length M, set the number of samples for smoothing processing, effectively reduce the output flicker, and fix the total driving current I allocated to the dual-color channel tot , as the global upper limit constraint of lighting intensity, provides the basis for light distribution ratio;

[0091] Calculate the effective duration of sunshine D = t set -t rise If D≤0, it will prompt invalid sunshine duration, please reset it; calculate the duration of daily dynamic lighting color temperature change, if it is invalid, prevent the system from malfunctioning;

[0092] Calculate the number of sampling moments in a day Specify the number of dynamically adjusted moments in a day to facilitate subsequent array initialization and indexing.

[0093] Through centralized management and automatic verification of key lighting system parameters, the problems of traditional lighting system configuration processes being cumbersome, error-prone, and difficult to detect anomalies have been effectively resolved. Key parameters are verified for rationality before system deployment and operation to prevent lighting failures, excessive energy consumption, or reduced comfort levels caused by parameter conflicts or extreme settings. For example, the system can detect problems such as conflicts in daylight start and end time settings and color temperature range violations in advance, prompting users to make corrections and ensuring the logical consistency of subsequent control links. For large-scale, intelligent lighting scenarios, this mechanism improves the system's ease of use, maintainability, and safety, making lighting solutions more aligned with human rhythms and actual needs, and providing a solid foundation for automated, unattended operation.

[0094] The generation of a sampling time sequence based on the effective sunshine duration and the sampling interval specifically includes:

[0095] Generate all dimming sampling moments to ensure continuous and smooth dynamic color temperature;

[0096] For each sampling point n = {0, 1, ..., N-1} calculate t[n] = t rise +nΔt; where n is the sampling number; t[n] is the nth sampling moment; forming a specific timestamp for each sampling point to ensure that dimming and color adjustment are aligned with the natural circadian rhythm;

[0097] If t[N-1]>tset , then let t[N-1]=t set ; Avoid the last sampling moment from exceeding the sunshine range to prevent the color temperature from changing during the invalid period.

[0098] By automatically generating a sampling sequence, we address the issues of discontinuous lighting, rhythmic disruption, and complex maintenance caused by manually setting sampling time nodes. This allows lighting adjustment sampling points to intelligently cover the entire effective lighting range, precisely aligning them with circadian rhythms and the human circadian clock, effectively ensuring smooth, natural, and real-time color temperature adjustment. It also automatically prevents the last sampling point from exceeding the time interval within the effective range, preventing abnormal lighting operation and improving system robustness. This mechanism enhances the dynamic responsiveness and user experience of intelligent lighting, making color temperature transitions smoother and more natural, and providing a solid time reference for subsequent target color temperature calculation and light mixing control.

[0099] The method of calculating the target color temperature curve at each moment based on the sampling moment and in accordance with the preset diurnal variation law to obtain the dynamic color temperature target sequence of the lighting output within the full cycle specifically includes:

[0100] Determine the optimal target color temperature for each sampling moment to achieve scientific human-centered lighting;

[0101] Calculate the normalized day-night phase at the nth moment Map the sampling points within a day to the interval [0,π] to achieve a periodic description of the color temperature changes during the day and night;

[0102] Get the target output color temperature T at the nth moment tgt [n]:

[0103] T tgt [n] = T min +(T max -T min )sin(φ[n]); Calculate the physiological health color temperature curve value for each sampling point, so that the lighting is both in line with natural changes and scientific and reasonable.

[0104] By calculating the target color temperature in real time for each sampling moment based on circadian rhythms and preset curves, this system addresses the abrupt color temperature changes and inability of traditional lighting systems to meet human physiological needs. This enables lighting color temperature to continuously and gradually respond to changes in natural light or user health needs, scientifically simulating the natural light environment and helping to improve the physiological comfort and health of indoor occupants. Furthermore, the dynamic generation of the target color temperature facilitates subsequent algorithms for precise light mixing and dynamic drive current distribution, fundamentally improving the human-friendliness, scenario adaptability, and market competitiveness of intelligent lighting systems. This control concept based on dynamic target curves is the technical foundation of modern healthy lighting and smart building lighting control.

[0105] The method of defining a light mixing function based on the color temperature-current model of the dual-color channel and the target color temperature curve and solving the distribution ratio of the dual-color channel current by mathematical methods specifically includes:

[0106] Calculate the two-color mixing ratio to ensure that the output color temperature is strictly consistent with the target color temperature;

[0107] Assume that the proportion of warm white channel current is r, which directly determines the distribution of two-color light distribution at each moment and is the core parameter of color temperature adjustment; then the warm white channel current I w =r·I tot , cold white channel current I c =(1-r)·I tot ; Calculate the actual operating current of the channel in real time based on the allocation ratio, which directly affects the LED output color temperature;

[0108] Construct mixed light color temperature function: Among them, T mix (r) is the warm and cool mixed output color temperature; it expresses the total output color temperature of the system after dual-color mixing, allowing accurate reverse calculation of the distribution ratio;

[0109] Expand and sort to get Ar 2 +Br+C=0, the color temperature mixing problem is transformed into a mathematical root problem, which can be strictly solved by algebraic methods; where A=I tot (α w +α c ) is the coefficient of the quadratic term; B = b w -b c -2α c I tot is the coefficient of the first-order term; C=α c I tot +b c -T tgt [n] is a constant term; each term reflects the sensitivity and bias of the color temperature adjustment of the two-color channel, providing necessary conditions for solving r;

[0110] Get the discriminant Δ=B 2 -4AC; Determine the existence and distribution of equation solutions to ensure that the mixed light adjustment results have physical meaning;

[0111] If Δ≥0, we get Among them, r1 and r2 are two roots of the quadratic equation, representing the possible solutions of the current proportion of the warm white channel;

[0112] Select the first root that satisfies 0≤r≤1 in turn. If both roots are there, take r1;

[0113] Screen out the legal light distribution ratio to avoid negative or over-limit channel current;

[0114] If Δ<0 or both roots are not in [0,1], it degenerates into a linear approximation:

[0115] in, This is the color temperature of the warm white channel alone at full current; is the color temperature of the cold white channel at full current; r[n] is the current ratio of the warm white channel at the nth sampling point;

[0116] When the quadratic method has no solution or the value is extreme, the system color temperature can still be adjusted continuously;

[0117] like Then let r[n] = 0.5; in extreme cases, ensure dual color balance and prevent the denominator from being zero;

[0118] Bounds cutoff: Ensure that the light distribution ratio is physically valid and the current distribution will never go out of bounds.

[0119] Through rigorous mathematical modeling and precise calculation of mixing ratios, this system overcomes the problems of large adjustment errors, blind light distribution, and poor system stability inherent in existing dual-color LED mixing methods. Based on the known target color temperature and the calibration model for each channel, the optimal current distribution at each moment is inferred, improving the accuracy and physical consistency of lighting color temperature control. Fault-tolerant mechanisms such as automatic boundary verification, outlier detection, and switching to linear approximation ensure that lighting output always operates within physically and safely permitted ranges, avoiding color temperature drift or channel damage. This mechanism also facilitates subsequent algorithm expansion and multi-channel system applications, enhancing system adaptability, robustness, and scalability, and providing technical support for high-end intelligent lighting control.

[0120] The method of allocating the obtained current values ​​of each channel and smoothing the current data using a sliding average method specifically includes:

[0121] Prevent output color temperature jumps caused by sudden ratio changes and improve lighting comfort;

[0122] Original current distribution: I w [n]=r[n]I tot , I c [n]=(1-r[n]I tot ; Among them, I w [n] is the warm white channel current at the nth moment; I c [n] is the current of the cold white channel at the nth moment; it distributes the actual driving current of the two-color channels to ensure the overall brightness stability of the system;

[0123] Get the number of valid samples K participating in the average at the nth moment n =min(M,n+1); dynamically determine the sliding window length to ensure that no samples are lost during the system startup phase;

[0124] Calculate the smoothed current of the warm white channel at moment n

[0125] Calculate the smoothed current of the cold white channel at moment n

[0126] Smooths raw current data, eliminates flicker and sudden changes, and optimizes human eye perception and system reliability.

[0127] By using a sliding average method to process channel current distribution at each moment, this method solves the visual discomfort caused by sudden output changes, flickering, and other issues encountered during dynamic adjustment in traditional lighting systems. The sliding average algorithm effectively reduces short-term variations caused by sampling, calculation, or driver errors, resulting in a smoother, more natural, continuous, and comfortable lighting output, optimizing the human eye's visual experience. This method also helps extend LED device life, avoids component stress and failure caused by excessive current fluctuations, and improves the long-term reliability of the system. Smoothing also facilitates the subsequent integration of more complex human-centric lighting strategies to support the lighting needs of diverse smart scenarios.

[0128] Outputting the finally determined warm white and cool white channel driving current values ​​to the lighting driver at each sampling moment specifically includes:

[0129] The smooth current command calculated each time is written into the drive system in time sequence to form high-quality lighting;

[0130] At each sampling time t[n] write To the LED driver; directly drive the LED channel, realize automatic color temperature change, and ensure that the lighting output is synchronized with the calculation;

[0131] If writing fails, the error log is recorded and the last valid value is maintained. When the system encounters a communication failure or hardware error, the log is recorded and the last valid instruction is maintained to prevent sudden black screen or erroneous operation.

[0132] Through high-frequency timed output and anomaly self-adaptation mechanisms, the risks of global blackout and uncontrolled lighting caused by communication anomalies or driver failures in traditional lighting systems are eliminated. This not only achieves real-time synchronization between drive current commands and lighting hardware, ensuring high consistency between theoretical color temperature and actual output, but also enhances the system's robustness and automated maintenance capabilities through anomaly recording and fault-tolerance mechanisms. In the event of unexpected conditions such as write failures, the system automatically maintains the last valid output value, avoiding adverse user experiences or equipment damage caused by sudden output changes. This mechanism enhances the engineering practicality and controllability of subsequent operations and maintenance of intelligent lighting systems, making them suitable for long-term, stable operation in critical locations.

[0133] After the daily lighting cycle ends, the lighting power supply is automatically cut off and restarted the next day according to the established process, specifically including:

[0134] The system is closed-loop to ensure energy saving and long-term stability;

[0135] If the last sampling point t[N-1]≥t set , then write (0,0) to cut off the power of the lamp; automatically enter the energy-saving mode to reduce energy consumption and LED life loss;

[0136] The system enters low power sleep mode until the next day rise Automatically return to step S3 and start again; realize continuous operation of lighting rhythm without human intervention, which is convenient for long-term deployment.

[0137] By automatically shutting off the system daily and then resuming it the next day, the system overcomes the high energy consumption, manual operation requirements, and frequent forgetfulness associated with traditional systems. It automatically enters a low-power sleep mode during off-hours, reducing standby energy consumption and overall operating costs. Furthermore, it automatically resumes the lighting control process the next day, enabling unattended, continuous operation. This greatly facilitates efficient management in long-term deployment scenarios such as smart buildings, factories, and campuses. This closed-loop mechanism not only saves energy and reduces consumption, but also protects the lifespan of equipment components, improving the overall reliability and intelligence of the lighting system, and aligning with the development trends of green buildings and smart cities.

[0138] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0139] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as within the scope of protection of the present invention.

Claims

1. A lighting control method based on artificial intelligence, characterized in that: include: S1. Apply predefined driving currents to the warm white channel and the cool white channel of the dual-color LED, respectively, and establish a linear calibration model between the color temperature of each channel and the driving current based on the measured output color temperature; S2. Define and verify system parameters, including sunshine start time, sunshine end time, target color temperature upper and lower limits, sampling time interval, sliding average filter window length, and total driving current of the two-color channel. Also, perform rationality check on the effective sunshine duration and determine the number of sampling times. S3. Generate a sampling time sequence based on the effective sunshine duration and sampling interval; S4. Based on the sampling time, according to the preset diurnal variation pattern, the target color temperature curve at each moment is calculated to obtain the dynamic color temperature target sequence of the lighting output within the full cycle; S5. Based on the color temperature-current model of the two-color channel and the target color temperature curve, define a light mixing function, and solve the distribution ratio of the two-color channel current through a mathematical method; S6. Allocate the current values ​​of each channel obtained and smooth the current data using a sliding average method; S7. At each sampling moment, output the finally determined warm white and cool white channel driving current values ​​to the lighting driver; S8. After the daily lighting cycle ends, the lighting power supply is automatically cut off and restarted the next day according to the established process.

2. The artificial intelligence-based lighting control method according to claim 1, characterized in that: The method of applying a predefined driving current to the warm white channel and the cool white channel of the dual-color LED and establishing a linear calibration model between the color temperature of each channel and the driving current according to the measured output color temperature specifically includes: Two different drive currents are applied to the warm white LED channel and the cool white LED channel respectively: Warm white channel current: I w1 , I w2 ; Corresponding measured color temperature: T w1 , T w2 ; Among them, I w1 I is the driving current of the first calibration point of the warm white channel; w2 is the driving current of the second calibration point of the warm white channel; T w1 For the warm white channel, the current I w1 Color temperature measured under T w2 For the warm white channel, the current I w2 Color temperature measured under Cold white channel current: I c1 , I c2 ; Corresponding measured color temperature: T c1 , T c2 ; Among them, I c1 I is the driving current of the first calibration point of the cold white channel; c2 is the driving current of the second calibration point of the cold white channel; T c1 For the cold white channel, the current I c1 Color temperature measured under T c2 For the cold white channel, the current I c2 Color temperature measured under Calculate the gain α of the warm white channel color temperature change on the current w and the linear bias term b of the warm white channel color temperature and current w : Calculate the color temperature change corresponding to the unit current change of the warm white channel, reflecting the sensitivity of the drive current adjustment to the color temperature; b. w =T w1 -α w I w1 ; Calculate the gain α of the cold white channel color temperature change on the current c and the linear bias term b of the cool white channel color temperature and current c : b c =T c1 -α c I c1 ; Construct arbitrary warm white channel current I w Output color temperature function T under w (I w ) and any cold white channel current I c Output color temperature function T under c (I c ): T w (I w )=α w I w +b w ,T c (I c )=α c I c +b c 。 3. The artificial intelligence-based lighting control method according to claim 2, characterized in that: The definition and verification of system parameters, the rationality check of the effective sunshine duration and the determination of the number of sampling moments specifically include: Preset by the designer: Sunshine start time t rise , end time of sunshine t set 、Target color temperature lower limit T min , target color temperature upper limit T max , sampling time interval Δt, sliding average filter window length M and the total driving current I fixedly allocated to the dual-color channel tot ; Calculate the effective duration of sunshine D = t set -t rise If D≤0, it will prompt invalid sunshine duration, please reset it; Calculate the number of sampling moments in a day 4. The artificial intelligence-based lighting control method according to claim 3, characterized in that: The generation of a sampling time sequence based on the effective sunshine duration and the sampling interval specifically includes: For each sampling point n = {0, 1, ..., N-1} calculate t[n] = t rise +nΔt; where n is the sampling number; t[n] is the nth sampling time; If t[N-1]>t set , then let t[N-1]=t set .

5. The artificial intelligence-based lighting control method according to claim 4, characterized in that: The method of calculating the target color temperature curve at each moment based on the sampling moment and in accordance with the preset diurnal variation law to obtain the dynamic color temperature target sequence of the lighting output within the full cycle specifically includes: Calculate the normalized day-night phase at the nth moment Get the target output color temperature T at the nth moment tgt [n]: T tgt [n]=T min +(T max -T min )sin(φ[n])。 6. The artificial intelligence-based lighting control method according to claim 5, characterized in that: The method of defining a light mixing function based on the color temperature-current model of the dual-color channel and the target color temperature curve and solving the distribution ratio of the dual-color channel current by mathematical methods specifically includes: Assuming the proportion of warm white channel current is r, the warm white channel current I w =r·I tot , cold white channel current I c =(1-r)·I tot ; Construct mixed light color temperature function: Among them, T mix (r) is the warm and cool mixed output color temperature; Expand and sort to get Ar 2 +Br+C=0;wherein, A=I tot (α w +α c is the coefficient of the quadratic term; B = b w -b c -2α c I tot is the coefficient of the first-order term; C=α c I tot +b c -T tgt [n] is a constant term; Get the discriminant Δ=B 2 -4AC; If Δ≥0, we get Among them, r1 and r2 are two roots of the quadratic equation, representing the possible solutions of the current proportion of the warm white channel; Select the first root that satisfies 0≤r≤1 in turn. If both roots are there, take r1; If Δ<0 or both roots are not in [0,1], it degenerates into a linear approximation: in, This is the color temperature of the warm white channel alone at full current; is the color temperature of the cold white channel at full current; r[n] is the current ratio of the warm white channel at the nth sampling point; like Then let r[n]=0.5; Bounds cutoff:

7. The artificial intelligence-based lighting control method according to claim 6, characterized in that: The method of allocating the obtained current values ​​of each channel and smoothing the current data using a sliding average method specifically includes: Original current distribution: I w [n]=r[n]I tot , I c [n]=(1-r[n])I tot ; Among them, I w [n] is the warm white channel current at the nth moment; I c [n] is the cold white channel current at the nth moment; Get the number of valid samples K participating in the average at the nth moment n =min(M,n+1); Calculate the smoothed current of the warm white channel at moment n Calculate the smoothed current of the cold white channel at moment n 8. The artificial intelligence-based lighting control method according to claim 7, characterized in that: Outputting the finally determined warm white and cool white channel driving current values ​​to the lighting driver at each sampling moment specifically includes: At each sampling time t[n] write To LED driver; If writing fails, the error log is recorded and the last valid value is maintained.

9. The artificial intelligence-based lighting control method according to claim 8, characterized in that: After the daily lighting cycle ends, the lighting power supply is automatically cut off and restarted the next day according to the established process, specifically including: If the last sampling point t[N-1]≥t set , then write (0,0) to turn off the power of the lamp; The system enters low power sleep mode until the next day rise Automatically returns to step S3 and starts again.

Images