A multi-dimensional micro-disturbance and space fusion LED light field control method and system

Abstract

The present application relates to the technical field of lighting control, and particularly relates to a multi-dimensional micro-disturbance and space fusion LED light field control method and system, comprising: performing a multi-dimensional joint micro-disturbance sequence containing brightness micro-disturbance and color temperature micro-disturbance on at least one lamp in a lighting system, and measuring total light tristimulus values in each disturbance state by at least one light sensor; based on the total light tristimulus values, constructing an overdetermined equation system for solving, and separating out the tristimulus values of the light component and the tristimulus values of the ambient light component of each lamp; based on the separated tristimulus values of the ambient light component and the preset target lighting parameters, calculating compensation control instructions of the lamps, and controlling the lamps to execute corresponding compensation control instructions. The present application solves the technical problems of inaccurate light component estimation, insufficient control precision and robustness caused by the limitation of separation means in the prior art.

Description

Technical Field

[0001] This invention relates to the field of lighting control technology, and in particular to a method and system for controlling LED light fields by multidimensional micro-perturbation and spatial integration. Background Technology

[0002] As intelligent lighting technology develops towards high-precision, multi-dimensional control, users' demands for the light environment have evolved from simple on / off control to precise control of brightness, color temperature, and color coordinates. However, in practical applications, the total light environment measured by light sensors is a mixture of controlled luminaire light and various uncontrolled ambient lights. To achieve accurate light environment control, it is essential to accurately separate the contributions of luminaire light and ambient light from the mixed measurements. Current mainstream methods typically rely on applying a single brightness perturbation to the luminaire and a very small number of measurements for estimation, leading to several inherent drawbacks: limited information dimension, inability to effectively decouple chromaticity information; insufficient measurement redundancy, resulting in poor algorithm robustness; and relatively coarse evaluation of luminaire correction capabilities. These shortcomings collectively make it difficult for existing technologies to achieve high-fidelity light field control in complex dynamic lighting environments. Therefore, there is an urgent need for an LED light field control method and system that can improve the accuracy and reliability of light component separation. Summary of the Invention

[0003] The main objective of this invention is to propose a method and system for controlling LED light field by multidimensional micro-perturbation and spatial fusion, aiming to solve the technical problems of inaccurate light component estimation, insufficient control precision and robustness caused by the limitations of separation methods in the prior art.

[0004] To achieve the above objectives, the first aspect of this invention proposes a method for controlling the LED light field by integrating multidimensional micro-perturbation and spatial fusion, comprising:

[0005] Step S100: Perform a multidimensional joint perturbation sequence including luminance micro-perturbation and color temperature micro-perturbation on at least one luminaire in the lighting system, and obtain the total light tristimulus value under each perturbation state by at least one light sensor;

[0006] Step S200: Based on the total light tristimulus values, construct an overdetermined system of equations to solve, and separate the tristimulus values ​​of the lamp light component and the ambient light component.

[0007] Step S300: Based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, calculate the compensation control command for the luminaire and control the luminaire to execute the corresponding compensation control command.

[0008] Preferably, step S100 includes:

[0009] Step S110: Generate a corresponding multidimensional joint micro-perturbation sequence according to the preset calibration mode. The sequence includes a reference measurement step and multiple brightness micro-perturbation steps and color temperature micro-perturbation steps arranged in a preset order. Each step has preset perturbation parameters and preset duration.

[0010] Step S120: Apply perturbations to one or more lamps sequentially according to the multidimensional joint micro-perturbation sequence;

[0011] Step S130: In each perturbation step, the total light tristimulus value under the current state is synchronously acquired by one or more optical sensors, and the average value is obtained by multiple samplings within the sampling window.

[0012] Preferably, step S200 includes:

[0013] Step S210: Based on the measured values ​​under each disturbance state, construct an overdetermined set of equations concerning the tristimulus values ​​of the luminaire;

[0014] Step S220: Solve the overdetermined system of equations using the weighted least squares method to obtain the optimal estimate of the tristimulus values ​​of the lamp.

[0015] Step S230: Based on the baseline measurement and the optimal estimate of the tristimulus values ​​of the luminaire light, calculate the tristimulus values ​​of the ambient light component.

[0016] Preferably, step S220 includes:

[0017] Step S221: Set a weight matrix according to the perturbation amplitude and signal-to-noise ratio of each perturbation step, wherein the measurement with the larger perturbation amplitude and the higher signal-to-noise ratio has a higher weight.

[0018] Step S222: Based on the weight matrix, use the weighted least squares method to solve for the optimal estimate of the light tristimulus values ​​of the lamp;

[0019] Step S223: Calculate the residuals of each measurement value and the standard deviation of all residuals. Determine the corresponding measurement values ​​whose absolute residual values ​​exceed the preset multiple of the standard deviation as outliers and remove them. Reconstruct and solve the overdetermined system of equations based on the removed measurement values.

[0020] Preferably, after step S230, the method further includes:

[0021] Step S240: For each sensor location, maintain a Kalman filter for real-time estimation and short-term prediction of the tristimulus values ​​of the ambient light component; the state vector of the Kalman filter at least contains the current tristimulus value estimate of the ambient light component at that sensor location and its rate of change.

[0022] Step S250: Establish and maintain a daylight pattern model, record the tristimulus values ​​of ambient light components at each sensor location during different time periods each day, and learn the diurnal variation pattern of ambient light by performing exponential weighted moving average processing on multi-day data.

[0023] Step S260: Using the daylight pattern model in conjunction with the Kalman filter includes: when the system starts up or is reinitialized, calling the daylight pattern model to provide an initial estimate of the ambient light component for the Kalman filter for the corresponding time period; during system operation, comparing the short-term prediction results of the Kalman filter with the long-term prediction reference values ​​of the daylight pattern model for the same time period, and if the deviation between the two exceeds a preset threshold, determining that the current illumination conditions are in an abnormal change state.

[0024] Preferably, step S300 includes:

[0025] Step S310: Construct a light color transfer matrix based on the tristimulus values ​​of the light components of each lamp. The light color transfer matrix is ​​used to describe the influence of the output changes of the tristimulus values ​​of each lamp on the measured values ​​of the tristimulus values ​​at each sensor position.

[0026] Step S320: Based on the light color transfer matrix, the tristimulus values ​​of the ambient light components, and the preset target lighting parameters, establish a quadratic programming optimization problem, where the decision variables are the output tristimulus values ​​of each lamp, the objective function is to minimize the total energy consumption of the system, and the constraints include physical model constraints on the tristimulus values ​​of each sensor location and lamp output range constraints.

[0027] Step S330: Solve the quadratic programming optimization problem using a quadratic programming solver to obtain a multi-lamp collaborative control command that optimizes the objective function under the constraints, and control each lamp to execute the collaborative control command.

[0028] Preferably, the method further includes the following after step S330:

[0029] Step S340: Use the S-curve function to generate a smooth transition curve for the lamp output from the current value to the target value. The expression is as follows:

[0030]

[0031] In the formula, The initial output value is for transition. The output value is the transition target, where t represents the current time starting from the transition start point. Let be the output value of the lamp at time t; The transition time is dynamically set according to the magnitude of output change; the greater the magnitude of change, the longer the transition time. This represents an S-shaped curve function; x represents the normalized time variable.

[0032] Step S350: Set a control dead zone. When the absolute value of the deviation between the total light tristimulus value at the sensor position and the target illumination parameter is less than the preset sensing threshold, maintain the current output of the lamp and do not perform any adjustment.

[0033] Step S360: Set the maximum allowable variation in brightness output and color temperature output of each lamp within a single control cycle to limit the output variation rate.

[0034] Preferably, the LED light field control method integrating multidimensional micro-perturbation and spatial fusion further includes an adaptive calibration step, specifically including: real-time monitoring of the change rate of the ambient light component, selecting the corresponding calibration mode according to the interval of the change rate, wherein the calibration modes are sequentially divided into a first calibration mode, a second calibration mode, and a third calibration mode according to the increasing calibration period; the first calibration mode, the second calibration mode, and the third calibration mode correspond to a multidimensional joint micro-perturbation sequence with increasing execution steps, finer perturbation amplitude gradient, or more comprehensive perturbation type combination, respectively; the mode switching adopts a hysteresis mechanism, wherein the change rate trigger threshold for switching from a mode with a longer calibration period to a mode with a shorter calibration period is higher than the change rate trigger threshold for reverse switching, and the switching from a mode with a longer calibration period to a mode with a shorter calibration period requires the change rate to be lower than the corresponding degradation threshold and to last for a preset time.

[0035] Preferably, the LED light field control method based on multidimensional micro-perturbation and spatial fusion is executed in a three-layer nested architecture comprising a first control layer, a second control layer, and a third control layer. The first control layer operates with a first control cycle and is used to perform tasks such as long-term trend modeling of ambient light components, time-series prediction based on Kalman filters, and updating the daylight pattern model. The second control layer operates with a second control cycle shorter than the first control cycle and is used to perform tasks such as multidimensional joint micro-perturbation sequence, light component separation, and updating the light color transfer matrix according to the mode determined by the adaptive calibration strategy. The third control layer operates with a third control cycle shorter than the second control cycle and, based on the output results of the first and second control layers, performs the construction and solution of a quadratic programming optimization problem to generate collaborative control commands, and finally sends the commands to each lamp in real time through smooth transition and anti-oscillation control steps.

[0036] A second aspect of this invention proposes a multi-dimensional micro-perturbation and spatial fusion LED light field control system, comprising:

[0037] A multidimensional joint micro-perturbation control module is used to perform a multidimensional joint micro-perturbation sequence including brightness micro-perturbation and color temperature micro-perturbation on at least one luminaire in a lighting system, and the total light tristimulus value under each perturbation state is measured by at least one light sensor.

[0038] The light component separation module is used to construct and solve an overdetermined system of equations based on the total light tristimulus values, and to separate the tristimulus values ​​of the lamp light component and the ambient light component.

[0039] The adaptive compensation control module is used to calculate the compensation control command of the lamp based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, and to control the lamp to execute the corresponding compensation control command.

[0040] The present invention provides a multi-dimensional micro-perturbation and spatial fusion LED light field control method and system. By introducing color temperature micro-perturbation and brightness micro-perturbation to form a multi-dimensional joint micro-perturbation sequence, it expands from a single brightness dimension to dual-dimensional detection of brightness and color temperature, improving the ability to accurately separate luminaire light and ambient light from mixed light. By constructing and solving an overdetermined system of equations based on the total light tristimulus values ​​under each perturbation state, and using multiple measurement redundancy for optimal estimation, it improves the accuracy of light component separation and robustness to sensor noise and sudden changes in ambient light. By comparing the separated ambient light component with the target value to generate compensation commands and smoothly execute them, it forms an active adaptive closed-loop control, improving the stability and fidelity of the light environment under different ambient light interferences.

[0041] Furthermore, this invention improves the standardization and repeatability of micro-perturbation detection by generating and executing a preset multi-dimensional joint micro-perturbation sequence in an orderly manner; improves the accuracy and stability of measurement data by averaging multiple samples synchronously within the sampling window; improves the reliability and noise resistance of light component estimation by constructing an overdetermined system of equations and solving it using the weighted least squares method; further improves estimation accuracy and robustness by introducing adaptive weights and outlier removal mechanisms; improves ambient light estimation accuracy and response timeliness by co-predicting with Kalman filtering and a daylight pattern model; enhances system startup efficiency and robustness by combining diurnal variation law learning and anomaly detection; improves multi-lamp collaborative sensing capability by constructing a light color transfer matrix to achieve spatial light field modeling; improves control energy efficiency and multi-objective satisfaction by global optimization through quadratic programming; improves user visual comfort by smoothing the transition with an S-curve; improves system stability and robustness by setting control dead zones and rate of change limits; improves environmental adaptability by dynamically selecting calibration modes and hysteresis switching mechanisms; and improves system efficiency and real-time response capability by executing in parallel using a three-layer nested architecture.

[0042] In summary, the LED light field control method and system proposed in this invention, which integrates multidimensional micro-perturbation and spatial fusion, solves the technical problems of inaccurate light component estimation, insufficient control precision and robustness caused by the limitations of separation methods in the prior art. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, 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 the structures shown in these drawings without creative effort.

[0044] Figure 1 A flowchart of an LED light field control method based on multidimensional micro-perturbation and spatial fusion provided in an embodiment of the present invention;

[0045] Figure 2 This is a flowchart of a multidimensional joint micro-perturbation sequence execution method provided in an embodiment of the present invention;

[0046] Figure 3 This is a flowchart of a method for solving the separation of luminaire light and ambient light components according to an embodiment of the present invention;

[0047] Figure 4 A flowchart of a method for solving and handling outliers using weighted least squares according to an embodiment of the present invention;

[0048] Figure 5 A flowchart illustrating a method for predicting ambient light based on a Kalman filter and a daylight pattern model, provided in an embodiment of the present invention.

[0049] Figure 6 This is a flowchart of a method for optimizing multi-lamp collaborative control based on light color transfer matrix and quadratic programming according to an embodiment of the present invention.

[0050] Figure 7 This is a flowchart of a method for smoothing transition and anti-oscillation control of lamp output provided in an embodiment of the present invention;

[0051] Figure 8 This is a schematic diagram of an LED light field control system that integrates multidimensional micro-perturbation and spatial fusion, provided in an embodiment of the present invention.

[0052] Figure 9 A schematic diagram of a computer device for controlling LED light field by multidimensional micro-perturbation and spatial fusion according to an embodiment of the present invention;

[0053] Figure 10The multidimensional joint micro-perturbation provided in one embodiment of the present invention is a sequence timing diagram;

[0054] Figure 11 This is a schematic diagram illustrating the effect of color temperature micro-perturbation on the CIE 1976 UCS chromaticity diagram according to an embodiment of the present invention;

[0055] Figure 12 This is a schematic diagram of multi-sensor spatial plaza fusion provided in an embodiment of the present invention;

[0056] Figure 13 This is a schematic diagram of a hierarchical control architecture and adaptive calibration strategy engine provided in an embodiment of the present invention;

[0057] Figure 14 This is a schematic diagram of a single-lamp, single-sensor office scenario according to an embodiment of the present invention;

[0058] Figure 15 This is a schematic diagram of a multi-light fixture and multi-sensor exhibition hall scene provided in an embodiment of the present invention.

[0059] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0060] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0061] It should be noted that if the embodiments of the present invention involve directional indicators, such as up, down, left, right, front, back, etc., the directional indicators are only used to explain the relative positional relationship and movement of the components in a specific posture. If the specific posture changes, the directional indicators will also change accordingly.

[0062] Furthermore, if the embodiments of this invention involve descriptions such as "first" or "second," these descriptions are for descriptive purposes only and should not be construed as indicating or implying their relative importance or implicitly specifying the number of technical features indicated. Therefore, a feature defined with "first" or "second" may explicitly or implicitly include at least one of those features. Additionally, the meaning of "and / or" throughout the text includes three parallel solutions; for example, "A and / or B" includes solution A, solution B, or a solution where both A and B are satisfied simultaneously. Furthermore, the technical solutions of the various embodiments can be combined with each other, but this must be based on the ability of those skilled in the art to implement them. When the combination of technical solutions is contradictory or impossible to implement, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection claimed by this invention.

[0063] like Figures 1 to 15 As shown, the first aspect of this invention proposes a method for controlling the LED light field by integrating multidimensional micro-perturbations and spatial fusion, comprising:

[0064] Step S100: Perform a multidimensional joint perturbation sequence including luminance micro-perturbation and color temperature micro-perturbation on at least one luminaire in the lighting system, and obtain the total light tristimulus value under each perturbation state by at least one light sensor;

[0065] Step S200: Based on the total light tristimulus values, construct an overdetermined system of equations to solve, and separate the tristimulus values ​​of the lamp light component and the ambient light component.

[0066] Step S300: Based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, calculate the compensation control command for the luminaire and control the luminaire to execute the corresponding compensation control command.

[0067] For details, see Figure 1In a specific embodiment of the present invention, a system comprising an intelligent control core module, an intelligent lighting network, and a light sensor network is taken as an example. The light sensor network consists of one or more (S, S≥1) light sensors, distributed in key locations in the space that need to be controlled, such as above an office desk or on the surface of an exhibit. Each sensor supports CIE XYZ tristimulus output and has a sampling rate of not less than 10Hz to ensure that stable measurement values ​​can be obtained within the short window of micro-disturbance. The sensors are connected to the intelligent control core module via wired (e.g., I2C, SPI bus) or wireless (e.g., Zigbee 3.0, BLE Mesh) communication protocols. The intelligent lighting network consists of one or more (L, L≥1) intelligent lighting fixtures distributed across various lighting locations in a space, such as ceilings, walls, and tracks. Each fixture supports independent adjustment of brightness (0-100% continuously adjustable) and color temperature (typically 2700K-6500K continuously adjustable). The fixtures connect to the control core of the intelligent control module via DALI-2 (Digital Addressable Lighting Interface), DMX512, or wireless protocols, enabling a millisecond-level response to control commands. The brightness and color temperature adjustment accuracy of the fixtures meets the requirements of a brightness resolution of no less than 0.1% and a color temperature resolution of no less than 10K. The intelligent control core module employs an ARM Cortex-A series processor or an equivalent SoC / edge computing device, running the aforementioned multi-dimensional micro-perturbation and spatial fusion LED light field control method. During system operation, the intelligent control core module periodically initiates the control method. First, it generates a standard multidimensional joint micro-perturbation sequence, which includes a baseline measurement step, at least one brightness reduction step, at least one brightness increase step, at least one color temperature warming adjustment step, and at least one color temperature cooling adjustment step. The intelligent control core module sends this sequence command to the intelligent luminaire to be calibrated. The luminaire sequentially switches to each perturbation state defined in the sequence, maintaining each state for a preset short time. Simultaneously, it controls one or more light sensors deployed at key spatial locations to perform multiple samplings during the stabilization period of each perturbation state, and transmits the calculated average total light tristimulus value back to the intelligent control core module in real time. After executing all steps of the micro-perturbation sequence, for each sensor location, the core module uses the corresponding multiple sets of total light tristimulus values ​​to construct an overdetermined system of equations with the tristimulus value contributed by the luminaire at that location as the unknown. This overdetermined system of equations is then solved to calculate the optimal estimated solution for the luminaire's light tristimulus value. The core module then subtracts the lamp light estimate from the baseline measurement at that location, thereby accurately separating the tristimulus values ​​of the ambient light component at that location.After completing the light component separation at all locations, the core module compares the ambient light tristimulus values ​​corresponding to each sensor location with the target tristimulus values ​​converted from the user-preset target lighting parameters for that area. By calculating the difference between the target value and the estimated ambient light value, the core module obtains the amount of tristimulus values ​​that the luminaires need to supplement to compensate for the influence of ambient light, and then converts this into brightness and color temperature adjustment commands for the corresponding luminaires. Finally, these compensation control commands are sent to the corresponding smart luminaires. Upon receiving the commands, the luminaires adjust their light output smoothly, making the ambient light measurement values ​​at each sensor location approach the user-set target, thus completing a full adaptive control loop.

[0068] Understandably, this embodiment introduces color temperature micro-perturbation and brightness micro-perturbation to form a multi-dimensional joint micro-perturbation sequence, expanding from a single brightness dimension to dual-dimensional detection of brightness and color temperature, thereby improving the ability to accurately separate luminaire light and ambient light from mixed light; by constructing and solving an overdetermined system of equations based on the total light tristimulus values ​​under each perturbation state, and utilizing multiple measurement redundancies for optimal estimation, the accuracy of light component separation and robustness to sensor noise and sudden changes in ambient light are improved; by comparing the separated ambient light component with the target value to generate compensation commands and smoothly executing them, an active adaptive closed-loop control is formed, improving the stability and fidelity of the light environment under different ambient light interferences.

[0069] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the number of steps, perturbation amplitude, or sequence of the multidimensional joint micro-perturbation sequence can be dynamically adjusted according to the lamp model or historical calibration effect to create an adaptive sequence that better matches the response characteristics of different hardware; or the amplitude of the brightness perturbation in the micro-perturbation sequence can be dynamically adjusted according to the ambient light intensity, using a larger perturbation amplitude when the ambient light is strong to improve the signal-to-noise ratio, and using a smaller perturbation amplitude when the ambient light is weak to reduce interference to the human eye.

[0070] Preferably, step S100 includes:

[0071] Step S110: Generate a corresponding multidimensional joint micro-perturbation sequence according to the preset calibration mode. The sequence includes a reference measurement step and multiple brightness micro-perturbation steps and color temperature micro-perturbation steps arranged in a preset order. Each step has preset perturbation parameters and preset duration.

[0072] Step S120: Apply perturbations to one or more lamps sequentially according to the multidimensional joint micro-perturbation sequence;

[0073] Step S130: In each perturbation step, the total light tristimulus value under the current state is synchronously acquired by one or more optical sensors, and the average value is obtained by multiple samplings within the sampling window.

[0074] For details, see Figure 2 , Figure 10 and Figure 11 In one specific embodiment of the present invention, the intelligent control core module generates a corresponding multi-dimensional joint micro-perturbation sequence according to a preset calibration mode, introducing color temperature micro-perturbation into the detection sequence, forming two orthogonal information dimensions with the traditional brightness micro-perturbation. See also Figure 11 When a luminance perturbation is performed, the luminaire's spectral power distribution (SPD) is scaled proportionally, and its CIE XYZ tristimulus values ​​change proportionally. This means that the vector of tristimulus value changes caused by the luminance perturbation is in the same direction as the original tristimulus values, providing only one dimension of information. On the CIE 1976 UCS chromaticity diagram, luminance perturbation does not change the chromaticity coordinates. When a color temperature perturbation is performed, if the luminaire's color temperature changes (e.g., from 4000K to 4100K) while the total luminous flux remains constant, the luminaire's spectral power distribution undergoes a shape change (enhancement of warm or cool tones), and its CIE XYZ tristimulus values ​​change non-proportionally. This means that the vector of tristimulus value changes caused by the color temperature perturbation is in a different direction than the original tristimulus values, providing a new dimension of information independent of luminance perturbation. On the CIE 1976 UCS chromaticity diagram, color temperature perturbation shifts the chromaticity coordinates along the Planckian Locus. See also Figure 10 In this embodiment, the multidimensional joint micro-perturbation sequence includes five steps executed sequentially, as shown in Table 1:

[0075] Table 1. Steps for the Standard Sequence of Multidimensional Joint Micro-Perturbation

[0076]

[0077] The first step is a baseline measurement, controlling the luminaire to maintain its current output. The second step is a brightness reduction step, for example, controlling the luminaire's brightness to decrease to 90% of its original value. The third step is a brightness increase step, for example, controlling the luminaire's brightness to increase to 105% of its original value. The fourth step is a color temperature warming step, for example, controlling the luminaire's color temperature to increase by 100K. The fifth step is a color temperature cooling step, for example, controlling the luminaire's color temperature to decrease by 100K. Each step has a preset duration, which is 100 milliseconds in this embodiment. The core module sends precise control commands for each step to the intelligent luminaire in this order. After each step command is executed, when the luminaire's output enters a stable phase, the core module sends a synchronous acquisition command to all light sensors. After receiving the command, each light sensor performs multiple rapid samplings of the current light environment within a preset sampling window, for example, 50 milliseconds, at a rate of not less than 10Hz. Then, all the original tristimulus values ​​obtained from the sampling are calculated arithmetically either locally or uploaded to the core module. Finally, the average total tristimulus value representing the light environment at each location in space under this disturbance step is obtained and recorded. The entire sequence can be executed within hundreds of milliseconds, far below the human eye's perception threshold for light changes, ensuring that the calibration process is imperceptible to the user and providing multiple sets of spatial light field data containing different perturbation information for subsequent calculations of the core module.

[0078] Understandably, this embodiment designs and executes a standard sequence containing two types of perturbations based on the physical principle that brightness and color temperature perturbations provide orthogonal information dimensions in the spectral and tristimulus value vector spaces. This provides a complete data foundation for the subsequent accurate calculation of light components in the three-dimensional color space. By setting conditions to keep the total luminous flux basically unchanged for the color temperature perturbation step in the sequence, it ensures that the color temperature perturbation vector carries independent chromaticity information to the maximum extent, rather than a mixture of brightness and chromaticity information, thereby improving the purity of the information dimension and the effectiveness of decoupling.

[0079] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the condition of keeping the total luminous flux basically unchanged in the color temperature perturbation can be replaced with keeping the lamp input power constant or other approximate conditions to simplify the control logic; or the sequence steps of completely separating the brightness perturbation and color temperature perturbation can be replaced with a step that allows small-amplitude brightness and color temperature composite perturbations to be applied simultaneously in a single step to explore more efficient information acquisition methods; or the analysis and decoupling method based on the CIE XYZ tristimulus value vector space can be replaced with a method of modeling and calculating based on similar orthogonal principles in other color spaces (such as CIE Lab) to adapt to different application standards or evaluation systems.

[0080] Preferably, step S200 includes:

[0081] Step S210: Based on the measured values ​​under each disturbance state, construct an overdetermined set of equations concerning the tristimulus values ​​of the luminaire;

[0082] Step S220: Solve the overdetermined system of equations using the weighted least squares method to obtain the optimal estimate of the tristimulus values ​​of the lamp.

[0083] Step S230: Based on the baseline measurement and the optimal estimate of the tristimulus values ​​of the luminaire light, calculate the tristimulus values ​​of the ambient light component.

[0084] For details, see Figure 3 In a specific embodiment of the present invention, after obtaining all the measurement values ​​corresponding to the micro-perturbation sequence consisting of N steps (including a baseline step M0 and N-1 perturbation steps), the intelligent control core module performs optical component separation calculation based on the following core physical model: the total light tristimulus value M measured by the sensor in any state is the sum of the tristimulus output value L of the lamp light and the tristimulus contribution value E of the ambient light in that state, i.e., M = L + E, and it is assumed that E remains constant in a fast calibration sequence. For each perturbation step... The total light tristimulus value measured by the sensor is:

[0085]

[0086] in For lighting fixtures in a disturbed state The output of the tristimulus values ​​under the following conditions The values ​​are the tristimulus values ​​of ambient light (assuming they remain constant over a short period of time throughout the calibration sequence).

[0087] The tristimulus output of the luminaire under disturbance state k can be expressed as:

[0088]

[0089] in, Let K be the tristimulus value of the luminaire under perturbation state k. These are the tristimulus values ​​of the luminaire under reference conditions. This is the brightness scaling factor (known, set by the control core). This is the color temperature offset factor (known and set by the control core). The vector of tristimulus value changes caused by a unit color temperature change (which can be obtained through lamp calibration or theoretical calculation).

[0090] benchmark measurement Substituting these values, we obtain the change in disturbance relative to the baseline for each step:

[0091]

[0092] For the X channel (similar for Y and Z channels):

[0093]

[0094] In the formula, This represents the change in the total optical tristimulus value measured by the sensor relative to the baseline state at the kth micro-perturbation step; express The X-channel component of the vector; This represents the total photostimulation value measured by the sensor under baseline conditions; This represents the vector of light output tristimulus values ​​under the reference state; express The X-channel component of the vector; express The X-channel component of the vector.

[0095] When there are N perturbations, N equations and 3 unknowns can be constructed. The overdetermined system of equations (overdetermined when N>3):

[0096]

[0097] Where A is an N×3 coefficient matrix, Let be the vector of the tristimulus values ​​of the lamp light to be solved, and b be the N×1 measurement difference vector.

[0098] Then, the intelligent control core module employs weighted least squares to independently solve the overdetermined equations constructed for the X, Y, and Z channels. By assigning weights to each equation (usually based on the signal-to-noise ratio or perturbation amplitude of the measurement), the weighted least squares method balances the reliability of different data during the solution process, ultimately yielding the tristimulus values ​​for the lamp's light reference. The statistically optimal estimate is obtained. Then, the tristimulus contribution E of the ambient light can be determined by simple subtraction: This allows for the separation of the tristimulus values ​​of the ambient light component from the mixed light measurements.

[0099] As an example, the above process describes a light component separation method for a single luminaire and a single sensor. When multiple luminaires and multiple sensors are present, the logic for light component separation needs to be extended accordingly. In a multi-luminaire, multi-sensor system, the total tristimulus value measurement result at each sensor location is the superposition of the light contributions from all luminaires and the ambient light contribution. In this case, the baseline measurement value of a single sensor is equal to the sum of the tristimulus contributions of each luminaire under the baseline state plus the ambient light contribution E. The baseline measurement equation should be extended to... ,in This represents the sum of the outputs of all L luminaires under the baseline state. Accordingly, the measurement equation for each perturbation step k becomes... , where L This is the output of luminaire j (j=1,2,…,L) under perturbation state k. To solve for the light contribution of each luminaire independently (i.e., to solve for each…)… If this is the case, a more complex system of equations needs to be constructed, with the number of unknowns proportional to the number of luminaires. Accordingly, during the time-division calibration process, when a micro-perturbation is applied to each luminaire, all sensors synchronously record the response. The constructed illumination transfer matrix is ​​used to describe the influence of the output changes of each luminaire on the measured values ​​of each sensor. The ambient light component is no longer obtained through simple subtraction, but by solving the global overdetermined system of equations to obtain the light contribution of all luminaires and the ambient light contribution simultaneously. Specifically, the system incorporates the ambient light tristimulus values ​​at each sensor location as unknowns into the system of equations and solves them together with the light tristimulus values ​​of each luminaire, thereby achieving the separation of the ambient light component from the light component of each luminaire in a multi-luminaire scenario.

[0100] Understandably, this embodiment introduces multiple measurement redundancies by constructing an overdetermined set of equations, and improves information utilization by using more equations than unknowns, thereby improving the reliability of the solution and the noise resistance. By subtracting the lamp light estimate from the reference measurement value to obtain the ambient light component, the purpose of accurately separating the ambient light from the mixed light is achieved, providing accurate ambient light information for subsequent compensation control.

[0101] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the solution method for the overdetermined equation system can be replaced with other robust estimation algorithms, such as iterative reweighted least squares or Huber estimation, to enhance the resistance to measurement outliers; or real-time weight calculation can be performed according to the system hardware resources, and a preset fixed weight can be used to reduce computational overhead when computational resources are limited; or the number of perturbation steps used in constructing the overdetermined equation system can be dynamically adjusted according to the quality of the actual measurement data, reducing the number of perturbation steps to speed up the calibration when the measurement consistency is good, and increasing the number of perturbation steps to obtain a better estimate when the measurement noise is large.

[0102] Preferably, step S220 includes:

[0103] Step S221: Set a weight matrix according to the perturbation amplitude and signal-to-noise ratio of each perturbation step, wherein the measurement with the larger perturbation amplitude and the higher signal-to-noise ratio has a higher weight.

[0104] Step S222: Based on the weight matrix, use the weighted least squares method to solve for the optimal estimate of the light tristimulus values ​​of the lamp;

[0105] Step S223: Calculate the residuals of each measurement value and the standard deviation of all residuals. Determine the corresponding measurement values ​​whose absolute residual values ​​exceed the preset multiple of the standard deviation as outliers and remove them. Reconstruct and solve the overdetermined system of equations based on the removed measurement values.

[0106] For details, see Figure 4 In a specific embodiment of the present invention, the weighted least squares (WLS) method is used to solve for the optimal estimate. Since overdetermined equation systems typically lack exact solutions, a weighting coefficient is first calculated and assigned to each equation in the system, corresponding to the measurement value at each micro-perturbation step, to form a diagonal weight matrix W. The calculation of the weighting coefficient follows two principles: first, it is proportional to the square of the amplitude of the corresponding perturbation step, because large perturbations generate significant signal changes and higher signal-to-noise ratios; second, it is inversely proportional to the variance estimate of multiple samples within the sampling window, to reduce the impact of high-noise measurement data. The specific weighting calculation formula is as follows:

[0107]

[0108] In the formula, Let W be the weight value assigned to the measurement equation corresponding to the k-th micro-perturbation step in the weight matrix W. The measured change caused by the perturbation at step k. The model, This is an estimate of the noise standard deviation for the k-th step measurement.

[0109] By setting weights using the above formula, steps with strong signals and low measurement noise have a greater impact on the final result during the solution process. This guides the weighted least squares method to obtain a statistically superior solution that is less sensitive to noise and interference in estimating the tristimulus values ​​of the lamp light.

[0110]

[0111] In the formula, This represents the optimal estimate of the tristimulus vector of the luminaire light under the reference state. The coefficient matrix, composed of known perturbation parameters, reflects the linear influence of each perturbation step on each component of the lamp's tristimulus values. It is an N×1 measurement difference vector, where each element corresponds to the change in the total optical tristimulus value measured under a perturbation step relative to the reference state; Given an N×N diagonal weight matrix, the elements on the diagonal are... The weights are determined by the corresponding weighting formulas and are used to differentiate the reliability of the measurement equations for each perturbation step in the solution process.

[0112] Based on this weight matrix, the core module uses the weighted least squares method to obtain an initial optimal estimate of the tristimulus values ​​of the lamp light, and then calculates the residuals of each equation. :

[0113]

[0114] In the formula, This represents the residual at the k-th perturbation step, which is the deviation between the actual measured change at that step and the model prediction based on the current estimate. The vector representing the k-th row of the coefficient matrix A corresponds to the influence coefficients of the brightness scaling factor and color temperature shift factor on each component of the lamp's tristimulus values ​​in the k-th perturbation step.

[0115] If the absolute value of a residual exceeds the standard deviation of all residuals 2.5 times (i.e.) If the value is not found, the measurement is determined to be an outlier (possibly caused by a sudden change in ambient light or transient interference from the sensor), and it is removed from the equation set and the solution is recalculated.

[0116] The tristimulus values ​​of the lamp were obtained by solving the problem. Then, the tristimulus values ​​of ambient light were obtained through simple subtraction:

[0117]

[0118] In the formula, This represents the obtained ambient light tristimulus value vector, which includes the components of ambient light in the X, Y, and Z channels; This represents the optimal estimation vector of the tristimulus values ​​of the luminaire obtained by solving, which includes the tristimulus value components of the X, Y, and Z channels of the luminaire under the baseline state; , , These are the components of the reference measurement value in the X, Y, and Z channels, respectively; , , These are the optimal estimates of the obtained tristimulus values ​​of the lamp light in the X, Y, and Z channels, respectively.

[0119] Understandably, this embodiment introduces an adaptive weight matrix based on the perturbation signal strength and measurement noise variance into the weighted least squares method. This gives high signal-to-noise ratio data a greater influence in the solution process, suppresses interference from low-quality data, and improves the statistical optimality and accuracy of estimating the light contribution of lamps under the inherent noise environment of sensors. By calculating the residuals after solving and performing outlier detection and removal based on the statistical standard deviation, it can automatically identify and eliminate outlier data points caused by instantaneous changes in ambient light or sensor disturbances, preventing them from contaminating the final solution results and improving the fault tolerance and robustness of the algorithm in dynamic real-world scenarios. By resolving the overdetermined equations after removing outliers, it ensures that the final estimation results are not affected by outlier data, further improving the accuracy and reliability of light component separation.

[0120] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the weight calculation formula can be replaced with other forms of signal-to-noise ratio evaluation functions, such as a linear combination of disturbance amplitude and noise variance, to adjust the trade-off between signal strength and noise sensitivity; or the fixed 2.5 times standard deviation as the outlier judgment threshold can be replaced with a threshold based on the t-distribution critical value or dynamically adjusted according to the dataset size, so that the judgment criteria are more statistically adaptable; or the process of weighted least squares combined with posterior outlier removal can be replaced with a one-step robust regression algorithm, such as using the Huber loss function or Tukey double weight function M estimation, to simultaneously achieve optimized estimation of Gaussian noise and resistance to gross errors in a single calculation.

[0121] Preferably, after step S230, the method further includes:

[0122] Step S240: For each sensor location, maintain a Kalman filter for real-time estimation and short-term prediction of the tristimulus values ​​of the ambient light component; the state vector of the Kalman filter at least contains the current tristimulus value estimate of the ambient light component at that sensor location and its rate of change.

[0123] Step S250: Establish and maintain a daylight pattern model, record the tristimulus values ​​of ambient light components at each sensor location during different time periods each day, and learn the diurnal variation pattern of ambient light by performing exponential weighted moving average processing on multi-day data.

[0124] Step S260: Using the daylight pattern model in conjunction with the Kalman filter includes: when the system starts up or is reinitialized, calling the daylight pattern model to provide an initial estimate of the ambient light component for the Kalman filter for the corresponding time period; during system operation, comparing the short-term prediction results of the Kalman filter with the long-term prediction reference values ​​of the daylight pattern model for the same time period, and if the deviation between the two exceeds a preset threshold, determining that the current illumination conditions are in an abnormal change state.

[0125] For details, see Figure 5 In a specific embodiment of the present invention, an independent Kalman filter is maintained for each sensor position i, and the state vector is... Defined as:

[0126]

[0127] In the formula, This represents the Kalman filter state vector at sensor position i at time t. This is the vector estimation of the tristimulus values ​​of the ambient light at sensor location i at time t. This is the vector estimate of the rate of change of the tristimulus values ​​of ambient light at sensor position i at time t.

[0128] Assuming the state changes at a constant rate, the transition model is expressed as follows:

[0129]

[0130] In the formula, This is the state transition matrix, used to predict the current state based on the state at the previous time step; This is the process noise vector, used to characterize the unmodeled dynamic uncertainties in the state transition model; This represents the predicted state vector value at time t+1; The time interval between two consecutive optical component separation calibration operations; The process noise is modeled as zero-mean Gaussian white noise with a covariance matrix of Q.

[0131] The observation model is represented as follows:

[0132]

[0133] In the formula, The ambient light estimate obtained for calibration; This is the observation matrix, used to map the state vector to observable quantities; The observation noise is represented by the covariance matrix R.

[0134] The prediction steps are as follows:

[0135]

[0136] In the formula, This represents the prior prediction of the state vector at time t+1 based on the state estimate at time t. This represents the posterior estimate of the state vector at time t based on all observations prior to time t. This represents the covariance matrix of the predicted state values ​​at time t+1. Let represent the covariance matrix of the posterior state estimate at time t; Q is the state transition matrix, used to describe the linear evolution of the state vector from the current time step to the next time step; Q is the process noise covariance matrix, used to quantify the uncertainty of the system model.

[0137] The update steps are as follows:

[0138]

[0139] In the formula, This is the Kalman gain matrix, used to weigh the weights of predicted and observed values ​​in state updates; This represents the posterior state estimate after fusing the observations at time t+1; This represents the ambient light estimate obtained through calibration at time t+1, which is the observation input of the Kalman filter; Let represent the covariance matrix of the posterior state estimate at time t+1; This represents the covariance matrix of the predicted state values ​​at time t+1. R is the identity matrix; R is the observation noise covariance matrix, used to quantify the uncertainty of sensor measurement noise.

[0140] After completing optical component separation in each calibration cycle, the intelligent control core module uses a Kalman filter to perform a prediction step. Based on the posterior state estimate at the current moment, it predicts the prior state value and covariance matrix of the ambient light component at the next moment. When the next calibration cycle arrives and new ambient light measurements are acquired, the intelligent control core module performs an update step, calculates the Kalman gain, and corrects the predicted state based on the observed values ​​to obtain the posterior state estimate, completing one full filtering iteration. Through this recursive method, the Kalman filter can continuously output the optimal estimate of the ambient light component and achieve short-term prediction of the ambient light state at future moments based on the rate of change component in the state vector.

[0141] Meanwhile, the intelligent control core module maintains a daylight pattern model, recording the tristimulus values ​​of ambient light components at different times of day for the light sensor in this office setting. An exponentially weighted moving average is used to process the multi-day data, increasing the influence of recent data on the model and gradually learning the diurnal variation curve of ambient light in this setting. Upon the system's first startup each day, the intelligent control core module uses historical data from the daylight pattern model for the current time period as the initial estimate of the Kalman filter's state vector, enabling the filter to quickly converge to an accurate state and avoiding estimation bias caused by a lack of observation data in the initial stage. During continuous system operation, the intelligent control core module compares the short-term prediction results of the Kalman filter with the long-term prediction reference values ​​of the daylight pattern model for the same time period. When the deviation exceeds a preset threshold, such as a sudden drop in ambient light due to cloud cover outside the window, the system determines that the current lighting conditions are in an abnormal state and can trigger a precision calibration mode or adjust the control strategy to quickly respond to environmental changes.

[0142] Understandably, this embodiment uses a Kalman filter to perform real-time estimation and short-term prediction of ambient light components. By utilizing a modeling method where the state vector includes the current value and its rate of change, it achieves accurate tracking of ambient light change trends, improving the system's ability to predict dynamic changes in ambient light and its response timeliness. By establishing a daylight pattern model and learning exponentially weighted moving averages from multi-day data, the system grasps the diurnal variation pattern of ambient light, providing accurate initial estimates for the Kalman filter, accelerating the filter convergence process, and reducing the adjustment time after system startup. By co-comparing the short-term predictions of the Kalman filter with the long-term predictions of the daylight pattern model, the system can identify abnormal light change states, thereby triggering corresponding adaptive calibration strategies and improving the system's robustness and intelligence in complex weather environments.

[0143] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the state transition model of the Kalman filter can be replaced from the assumption of uniform change to a uniformly accelerated model to adapt to complex scenarios where the rate of change of ambient light itself is also changing; or the exponentially weighted moving average decay factor of the daylight pattern model can be adaptively adjusted according to seasonal changes, and recent data can be given higher weights during seasonal transitions to speed up model updates; or the collaborative mechanism between the Kalman filter and the daylight pattern model can be extended to multi-model fusion prediction, introducing weather forecast data or information from other nodes in the light sensor network as auxiliary inputs to further improve the accuracy of ambient light prediction.

[0144] Preferably, step S300 includes:

[0145] Step S310: Construct a light color transfer matrix based on the tristimulus values ​​of the light components of each lamp. The light color transfer matrix is ​​used to describe the influence of the output changes of the tristimulus values ​​of each lamp on the measured values ​​of the tristimulus values ​​at each sensor position.

[0146] Step S320: Based on the light color transfer matrix, the tristimulus values ​​of the ambient light components, and the preset target lighting parameters, establish a quadratic programming optimization problem, where the decision variables are the output tristimulus values ​​of each lamp, the objective function is to minimize the total energy consumption of the system, and the constraints include physical model constraints on the tristimulus values ​​of each sensor location and lamp output range constraints.

[0147] Step S330: Solve the quadratic programming optimization problem using a quadratic programming solver to obtain a multi-lamp collaborative control command that optimizes the objective function under the constraints, and control each lamp to execute the collaborative control command.

[0148] For details, see Figure 6 and Figure 12 In a specific embodiment of the present invention, the system includes L luminaires and S sensors. The light color transfer matrix K is an S×L dimensional matrix, where each element K[i,j] represents the change in the measured tristimulus value caused by a unit change in the light output of luminaire j (characterized by tristimulus values) at sensor i. This matrix describes the influence of all luminaires on the tristimulus values ​​at all key locations in the system and serves as the mathematical basis for spatial perception and collaborative control. The system dynamically constructs this matrix using a time-sharing calibration strategy. The specific process is as follows: First, a baseline measurement is performed, controlling all luminaires to maintain their current output and commanding all sensors to sample synchronously, recording the baseline total light tristimulus value M at each sensor position i. 0i (i=1,2,...,S). Next, a per-lamp calibration is performed, sequentially applying an active micro-perturbation to each lamp j (j=1,2,...,L), ensuring that the output of all other lamps remains unchanged while perturbing lamp j. During this perturbation, all S sensors simultaneously measure and record the total light tristimulus value M after the perturbation. ji Then, matrix element calculations are performed. For each sensor i, the change in measurement value caused by the disturbance of lamp j is calculated. The output change applied to lamp j This is a tristimulus value vector. Based on this, the elements in the light color transfer matrix corresponding to sensor i and luminaire j can be calculated, i.e. The division here represents the generalized linear influence coefficient. The above lamp-by-lamp calibration process is repeated until all L lamps have undergone an independent perturbation and measurement, ultimately completing the construction of the entire light color transfer matrix K, thus providing an accurate spatial illumination model for subsequent global optimization calculations.

[0149] For ease of understanding, see appendix Figure 12 This diagram illustrates the basic concept of the light color transfer matrix K. Figure 12 The diagram uses a light transfer matrix as an example (demonstrating only light parameters). The figure visually represents the light influence relationship between each luminaire and each sensor location in S×L dimensional matrix form. This diagram focuses on showing the logical structure of spatial association. It should be noted that in actual high-fidelity colorimetry control applications, this matrix will be expanded to fully characterize the comprehensive influence of the tristimulus output values ​​(X, Y, Z) of each luminaire on the tristimulus measurement values ​​of each sensor location. The expansion principle is consistent with the spatial association logic shown in the figure, that is, establishing a more refined linear mapping model for each pair of luminaire-sensor relationships in the three-dimensional color space.

[0150] After the dynamic construction of the light and color transfer matrix is ​​completed, the system can use this matrix for global optimization and collaborative control. Given the light transfer matrix K and the predicted ambient light values ​​Ê at each sensor location... i and the target set by the user for each sensor location i In this case, the system establishes the following quadratic programming problem:

[0151] The decision variables are the output tristimulus vectors of each lamp. ,in This represents the tristimulus output of lamp j.

[0152] The objective function is a weighted sum of minimizing the total system energy consumption and the output smoothness, expressed as:

[0153]

[0154] In the formula, For lighting fixtures The power coefficient, Let λ be the output of lamp j in the previous cycle, and λ be the smoothness weighting factor used to suppress drastic changes in lamp output.

[0155] The constraints include:

[0156] The photophysical model constraint states that the total photoluminescence tristimulus value at each sensor location should equal the sum of the contributions from all luminaires and the ambient light, expressed as follows: Deviations are allowed within a preset tolerance range;

[0157] Color difference constraint: for each sensor position i, the color difference between its color coordinates and the target color coordinates. The color difference must not exceed the user-defined color difference threshold ε, expressed as follows: ;

[0158] Luminaire range constraint, i.e., the output tristimulus value L of each luminaire. jIt should meet the requirements within its physical adjustable range. ;

[0159] Rate of change constraint, that is, the change in the tristimulus output values ​​of each lamp between adjacent control cycles does not exceed the preset maximum allowable change. The expression is ;

[0160] The intelligent control core module solves the optimization problem using a quadratic programming solver (such as the interior point method or the activity set method), calculates the multi-lamp collaborative control command L* that optimizes the objective function under all constraints, and sends the command to each lamp for execution, thereby realizing the global optimal light field adaptive control in multi-sensor, multi-lamp scenarios.

[0161] In a specific embodiment of the present invention, the interior-point solver configured in the intelligent control core module transforms a quadratic programming problem with inequality constraints into a series of unconstrained or equality-constrained subproblems that gradually approach the optimal solution by introducing a logarithmic barrier function. The solver iterates from an initial interior point that strictly satisfies all inequality constraints. In each iteration, it constructs a Lagrangian function that integrates the original objective function with a barrier term representing the degree of constraint violation. Subsequently, the solver calculates the gradient of this Lagrangian function with respect to the decision variables (i.e., the tristimulus vector L of each lamp's output) and the Hessian matrix, applies Newton's method or a quasi-Newton method to determine the search direction and step size, thereby updating the solution vector and simultaneously adjusting a central path parameter to ensure the iterative path remains within the feasible region and gradually approaches the optimal solution on the boundary. This process continues until the change in the solution vector, the decrease in the objective function, or the duality gap is less than a preset convergence tolerance. The resulting solution L* is then considered the globally optimal cooperative control command that satisfies all constraints.

[0162] As an alternative implementation, the intelligent control core module can be configured with an active set method solver to proactively predict the constraints that will be effective at the optimal solution, thereby simplifying the original problem into a quadratic programming subproblem with equality constraints. The solver starts with an initial feasible solution, treats all current constraints as inactive, and solves only an unconstrained problem. Depending on whether the solution violates a constraint, it adds the violated constraint to a "working set" (i.e., the active set), and then resolves a subproblem subject only to equality constraints within that working set. After each subproblem is solved, it checks the sign of the corresponding Lagrange multipliers (i.e., dual variables). If a multiplier corresponding to an active constraint is negative, it means that the constraint may not be active at the optimal solution, and it is removed from the working set. Through this iterative cycle of "identifying the active set - solving the equality-constrained subproblem - updating the working set," the solver continuously corrects its estimation of the active set and updates the solution vector until it finds a solution L* that satisfies the Caro-Kuhn-Tucker (KKT) optimality condition, and the working set remains stable. The activity set method is highly efficient when there are many constraints but few actually activated at the optimal solution. It is suitable for scenarios in this invention where there may be a large number of physical range constraints and rate of change constraints for the lamps, but only a portion of them are effective at the boundary.

[0163] Understandably, this embodiment constructs a light color transfer matrix through a time-sharing calibration strategy, quantifies the linear mapping relationship between the output changes of each lamp and the measured values ​​of each sensor, and realizes the mathematical modeling of the spatial light field distribution in a multi-lamp, multi-sensor system, laying the physical model foundation for global collaborative control. By establishing a quadratic programming optimization problem with the goal of minimizing the total system energy consumption while taking into account the smoothness of the output, and incorporating photophysical model constraints, color difference constraints, lamp output range constraints, and rate of change constraints into the solution framework, global energy consumption is optimized while meeting the high-fidelity light environment target of each area, avoiding mutual interference and oscillation between lamps under traditional independent control methods. The optimization calculation is completed in milliseconds using efficient solvers such as the interior point method or the active set method, ensuring that the control system has real-time response capability and realizing rapid dynamic control of multi-lamp collaboration.

[0164] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the construction process of the light color transfer matrix can be extended from complete calibration to incremental update, and only perturbation calibration can be performed on lamps that have changed significantly or have not been calibrated for a long time, so as to improve the efficiency of matrix maintenance in large-scale systems; or the smoothness weight factor in the quadratic programming can be set differently according to the type of lamp, giving higher smoothness weight to lamps that are prone to visual flicker and lower weight to lamps with fast response to improve system sensitivity; or the quadratic programming solver can be adaptively selected according to the system computing resources, using the interior point method to obtain higher solution accuracy when computing power is sufficient, and using the active set method to balance solution speed and resource consumption when computing power is limited.

[0165] Preferably, the method further includes the following after step S330:

[0166] Step S340: Use the S-curve function to generate a smooth transition curve for the lamp output from the current value to the target value. The expression is as follows:

[0167]

[0168] In the formula, The initial output value is for transition. The output value is the transition target, where t represents the current time starting from the transition start point. Let be the output value of the lamp at time t; The transition time is dynamically set according to the magnitude of output change; the greater the magnitude of change, the longer the transition time. This represents an S-shaped curve function; x represents the normalized time variable.

[0169] Step S350: Set a control dead zone. When the absolute value of the deviation between the total light tristimulus value at the sensor position and the target illumination parameter is less than the preset sensing threshold, maintain the current output of the lamp and do not perform any adjustment.

[0170] Step S360: Set the maximum allowable variation in brightness output and color temperature output of each lamp within a single control cycle to limit the output variation rate.

[0171] For details, see Figure 7 In a specific embodiment of the present invention, after the intelligent control core module obtains a new set of multi-lamp collaborative control instructions L* through quadratic programming, it does not immediately send the instruction values ​​directly to the lamps. Instead, it initiates a smooth transition and anti-oscillation control process. For each lamp that needs adjustment, its current output value is first recorded as the transition start value. The target instruction value obtained from the optimization solution is used as the transition endpoint. Next, based on the magnitude of the difference between the target value and the current value, a transition time is dynamically calculated using a linear mapping function. The greater the range of change, the better. The longer; then according to the formula Generate a smooth control curve, where , It is an S-shaped curve function that satisfies S(0)=0, S(1)=1 and whose first and second derivatives are zero at the endpoints. Specific parameters can also be adjusted as needed. Finally, during the transition period, the instantaneous target value for each control cycle is calculated in real time according to this curve and sent to the lamps. At the same time, the anti-oscillation control mechanism is also running. The core module will continuously monitor the actual total light tristimulus values ​​at each sensor location and compare them with the preset target illumination parameters. When the absolute value of the deviation between the measured value at a certain location and the target value is less than the preset human eye perception threshold (e.g., brightness difference less than 5%, chromaticity difference less than 5%), the target illumination value is determined. When the value is less than 0.002, the control logic enters a "dead zone," and the core module maintains the current output of the luminaire, no longer executing new adjustment commands. This avoids meaningless oscillations and frequent adjustments near the target value caused by measurement noise or minor disturbances. Furthermore, maximum allowable variations within a single control cycle are set for the brightness and color temperature channels of each luminaire, such as brightness variations not exceeding 3% per cycle and color temperature variations not exceeding 30K per cycle. Any change in the instantaneous target value calculated from the smoothing curve relative to the output of the previous cycle is limited to this maximum variation, thus physically limiting the maximum rate of change of the output and further ensuring the smoothness of changes in the lighting environment.

[0172] Understandably, this embodiment uses an S-curve function to control the smooth transition of the lamp output from the current value to the new target value, making the change process of the light environment continuous, differentiable, and stepless, eliminating the flicker or sudden change that the human eye can perceive, and improving visual comfort and light environment quality. By setting a control dead zone, unnecessary adjustments are suppressed when the deviation between the light environment and the target is less than the human eye's perception threshold, effectively avoiding frequent fine-tuning of the lamp output and system oscillations caused by sensor noise or slight fluctuations in ambient light, thus improving system stability and energy efficiency. By setting single-cycle maximum change limits for the brightness and color temperature output of the lamps, the maximum rate of change of the output is constrained from the root, forming a double guarantee with the smooth transition curve, preventing output jumps caused by drastic changes in optimization instructions or system anomalies, further improving the system's reliability and anti-interference capability.

[0173] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the fixed S-curve function can be replaced with other forms of smooth transition functions, such as the normalized form of the hyperbolic tangent function, to achieve different transition dynamic characteristics; or the fixed human eye perception threshold in the control dead zone can be replaced with an adaptive threshold that is dynamically adjusted according to the ambient light level or time to adapt to the changes in visual sensitivity of the human eye under different background brightness.

[0174] Preferably, the LED light field control method integrating multidimensional micro-perturbation and spatial fusion further includes an adaptive calibration step, specifically including: real-time monitoring of the change rate of the ambient light component, selecting the corresponding calibration mode according to the interval of the change rate, wherein the calibration modes are sequentially divided into a first calibration mode, a second calibration mode, and a third calibration mode according to the increasing calibration period; the first calibration mode, the second calibration mode, and the third calibration mode correspond to a multidimensional joint micro-perturbation sequence with increasing execution steps, finer perturbation amplitude gradient, or more comprehensive perturbation type combination, respectively; the mode switching adopts a hysteresis mechanism, wherein the change rate trigger threshold for switching from a mode with a longer calibration period to a mode with a shorter calibration period is higher than the change rate trigger threshold for reverse switching, and the switching from a mode with a longer calibration period to a mode with a shorter calibration period requires the change rate to be lower than the corresponding degradation threshold and to last for a preset time.

[0175] For details, see Figure 13 In a specific embodiment of the present invention, the intelligent control core module incorporates an adaptive calibration strategy engine to select the most suitable calibration mode in real time based on the dynamic characteristics of ambient light. It continuously monitors the rate of change of ambient light components at each sensor location, estimated by a Kalman filter, and dynamically switches between three predefined calibration modes based on this rate of change: when the rate of change of ambient light is below a threshold T1 (e.g., 5 lux / min), the environment is considered stable, and the first calibration mode is activated, corresponding to a micro-perturbation sequence with a longer execution cycle and the simplest steps; when the rate of change is between thresholds T1 and T2 (e.g., 20 lux / min), the environment is considered to be changing normally, and the second calibration mode is activated, corresponding to a standard sequence with a moderate execution cycle and complexity; when the rate of change exceeds threshold T2, the environment is considered to be undergoing a drastic change, and the third calibration mode is activated, corresponding to a sequence with the shortest execution cycle, the most precise steps, and the most comprehensive information dimensions. The specific multi-dimensional joint micro-perturbation sequences corresponding to these three modes are shown in the table below.

[0176] The first calibration mode (i.e., the simplified mode in the figure) sequence is suitable for rapid calibration in stable environments, and its sequence is shown in Table 2:

[0177] Table 2. Steps for the micro-perturbation sequence in the first calibration mode (simplified mode)

[0178]

[0179] The second calibration mode (i.e., the standard mode in the figure) sequence is applicable to routine calibration, and its sequence is shown in Table 3:

[0180] Table 3. Steps for the micro-perturbation sequence in the second calibration mode (standard mode)

[0181]

[0182] The third calibration mode (i.e., the precision mode in the figure) sequence is suitable for high-precision requirements or when there are drastic environmental changes. Its sequence is shown in Table 4:

[0183] Table 4. Steps for the micro-perturbation sequence in the third calibration mode (precision mode)

[0184]

[0185] Mode switching employs a hysteresis mechanism to prevent jitter: the rate of change trigger threshold (5 lux / min) required to switch from the first calibration mode to the second calibration mode is higher than the threshold for reverse switching (3.5 lux / min); similarly, the threshold for upgrading to the third calibration mode (20 lux / min) is higher than the threshold for downgrading back to the second calibration mode (14 lux / min). Furthermore, when downgrading from a short-cycle, high-precision mode (such as the third calibration mode) to a long-cycle, low-precision mode (such as the second calibration mode), the downgrading is only performed after the ambient light rate of change has remained below the downgrading threshold for a preset time (such as 30 seconds). This ensures that even after a brief fluctuation and the environment stabilizes, the system can maintain a relatively high-precision calibration state for a period of time, thus achieving an intelligent balance between accuracy, response speed, and system overhead.

[0186] Understandably, this embodiment introduces an adaptive strategy that dynamically selects the calibration mode based on the ambient light change rate, enabling the system to adopt a simplified calibration with a long cycle and low overhead when the environment is stable, thereby saving energy consumption and computing resources and improving the overall efficiency and adaptability of the system in different dynamic scenarios. By designing a hysteresis mechanism for mode switching that includes an asymmetric threshold and a state holding time, it effectively avoids frequent oscillation switching of the calibration mode when the ambient light change rate fluctuates near the threshold, thereby improving the stability and reliability of the system's working state.

[0187] Based on the above technical solutions, those skilled in the art can make corresponding equivalent improvements according to the specific characteristics of the application scenario or system requirements. For example, the fixed rate of change threshold can be replaced with an adaptive threshold that is dynamically adjusted through online learning, so that the system can better adapt to the unique light change patterns of different installation locations; or the three discrete calibration modes can be replaced with calibration modes with more levels or parameters that can be continuously fine-tuned, so as to achieve more precise control over the trade-off between resources and accuracy.

[0188] Preferably, the LED light field control method based on multidimensional micro-perturbation and spatial fusion is executed in a three-layer nested architecture comprising a first control layer, a second control layer, and a third control layer. The first control layer operates with a first control cycle and is used to perform tasks such as long-term trend modeling of ambient light components, time-series prediction based on Kalman filters, and updating the daylight pattern model. The second control layer operates with a second control cycle shorter than the first control cycle and is used to perform tasks such as multidimensional joint micro-perturbation sequence, light component separation, and updating the light color transfer matrix according to the mode determined by the adaptive calibration strategy. The third control layer operates with a third control cycle shorter than the second control cycle and, based on the output results of the first and second control layers, performs the construction and solution of a quadratic programming optimization problem to generate collaborative control commands, and finally sends the commands to each lamp in real time through smooth transition and anti-oscillation control steps.

[0189] For details, see Figure 13 In a specific embodiment of the present invention, the intelligent control core module operates within a three-layer nested architecture comprising a first control layer, a second control layer, and a third control layer. The first control layer operates on a 60-second cycle, responsible for performing long-term trend modeling of the ambient light components. It performs time-series prediction of the ambient light tristimulus values ​​based on a Kalman filter and continuously updates the daylight pattern model, recording and learning the ambient light variation patterns at different times of the day to provide long-term trend references for the system. The second control layer operates on a 10-second cycle, executing a multi-dimensional joint micro-perturbation sequence according to the pattern determined by the adaptive calibration strategy. During the perturbation process, it separates the light components and updates the color transfer matrix to ensure the accuracy of the spatial illumination model. The third control layer operates on a 200-millisecond cycle, constructing and solving a quadratic programming optimization problem based on the ambient light prediction values ​​provided by the first control layer and the updated color transfer matrix from the second control layer. It generates multi-lamp collaborative control commands and sends these commands to the lamps in real time through smooth transition and anti-oscillation control steps. The three levels run in parallel and each has its own independent timing. When the execution cycle of a task in a certain level is reached, the corresponding task is triggered without waiting for the other levels to complete, thereby achieving decoupling and parallel processing of tasks at different time scales.

[0190] Understandably, this embodiment employs a three-layer nested control architecture to decouple tasks of different time scales and complexities, such as long-term learning and prediction of ambient light, online calibration and modeling of the system, and real-time optimization and control of the light field, into independent layers for execution. This improves the system's modularity, operational efficiency, and maintainability. By setting differentiated operating cycles for different control layers, time-consuming prediction and learning tasks can be executed smoothly in the background, while closed-loop control tasks requiring rapid response can run at millisecond-level frequencies, achieving an optimal balance between the rational allocation of computing resources and the real-time response of the system.

[0191] See Figure 8 As shown, the second aspect of this invention proposes a multi-dimensional micro-perturbation and spatial fusion LED light field control system, comprising:

[0192] A multidimensional joint micro-perturbation control module is used to perform a multidimensional joint micro-perturbation sequence including brightness micro-perturbation and color temperature micro-perturbation on at least one luminaire in a lighting system, and the total light tristimulus value under each perturbation state is measured by at least one light sensor.

[0193] The light component separation module is used to construct and solve an overdetermined system of equations based on the total light tristimulus values, and to separate the tristimulus values ​​of the lamp light component and the ambient light component.

[0194] The adaptive compensation control module is used to calculate the compensation control command of the lamp based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, and to control the lamp to execute the corresponding compensation control command.

[0195] A third aspect of the present invention also provides a storage medium storing a multi-dimensional micro-perturbation and spatial fusion LED light field control processing program, wherein when the multi-dimensional micro-perturbation and spatial fusion LED light field control program is executed by a processor, it implements the steps of the multi-dimensional micro-perturbation and spatial fusion LED light field control method as described in any of the above embodiments.

[0196] See Figure 9 As shown, the fourth aspect of the present invention provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the LED light field control method of multidimensional micro-perturbation and spatial fusion as described in any embodiment of the first aspect.

[0197] The processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. This processor is the control center for the multi-dimensional micro-perturbation and spatial fusion LED light field control system, connecting various parts of the operational device through various interfaces and lines.

[0198] The memory can be used to store the computer program and / or modules. The processor realizes various functions of the multi-dimensional micro-perturbation and spatial fusion LED light field control system by running or executing the computer program and / or modules stored in the memory, and by calling the data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system, at least one application program required for a function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created according to the use of the mobile phone (such as audio data, phone book, etc.). In addition, the memory may include high-speed random access memory, and may also include non-volatile memory, such as hard disk, memory, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0199] Compared with the prior art, the beneficial effects of the present invention include at least the following:

[0200] The present invention provides a multi-dimensional micro-perturbation and spatial fusion LED light field control method and system. By introducing color temperature micro-perturbation and brightness micro-perturbation to form a multi-dimensional joint micro-perturbation sequence, it expands from a single brightness dimension to dual-dimensional detection of brightness and color temperature, improving the ability to accurately separate luminaire light and ambient light from mixed light. By constructing and solving an overdetermined system of equations based on the total light tristimulus values ​​under each perturbation state, and using multiple measurement redundancy for optimal estimation, it improves the accuracy of light component separation and robustness to sensor noise and sudden changes in ambient light. By comparing the separated ambient light component with the target value to generate compensation commands and smoothly execute them, it forms an active adaptive closed-loop control, improving the stability and fidelity of the light environment under different ambient light interferences.

[0201] Furthermore, this invention improves the standardization and repeatability of micro-perturbation detection by generating and executing a preset multi-dimensional joint micro-perturbation sequence in an orderly manner; improves the accuracy and stability of measurement data by averaging multiple samples synchronously within the sampling window; improves the reliability and noise resistance of light component estimation by constructing an overdetermined system of equations and solving it using the weighted least squares method; further improves estimation accuracy and robustness by introducing adaptive weights and outlier removal mechanisms; improves ambient light estimation accuracy and response timeliness by co-predicting with Kalman filtering and a daylight pattern model; enhances system startup efficiency and robustness by combining diurnal variation law learning and anomaly detection; improves multi-lamp collaborative sensing capability by constructing a light color transfer matrix to achieve spatial light field modeling; improves control energy efficiency and multi-objective satisfaction by global optimization through quadratic programming; improves user visual comfort by smoothing the transition with an S-curve; improves system stability and robustness by setting control dead zones and rate of change limits; improves environmental adaptability by dynamically selecting calibration modes and hysteresis switching mechanisms; and improves system efficiency and real-time response capability by executing in parallel using a three-layer nested architecture.

[0202] In summary, the LED light field control method and system proposed in this invention, which integrates multidimensional micro-perturbation and spatial fusion, solves the technical problems of inaccurate light component estimation, insufficient control precision and robustness caused by the limitations of separation methods in the prior art.

[0203] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention. Equivalent structural transformations made using the description and drawings of the present invention, or direct / indirect applications in other related technical fields, are all included within the scope of patent protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0204] It should be noted that the embodiment implemented by the LED light field control system of multidimensional micro-perturbation and spatial fusion in this invention can be referred to in conjunction with the embodiment implemented by the LED light field control method of multidimensional micro-perturbation and spatial fusion, and will not be described in detail in this invention.

[0205] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.

Claims

1. A method for controlling the LED light field by integrating multidimensional micro-perturbation and spatial manipulation, characterized in that, include: Step S100: Perform a multidimensional joint perturbation sequence including luminance micro-perturbation and color temperature micro-perturbation on at least one luminaire in the lighting system, and obtain the total light tristimulus value under each perturbation state by at least one light sensor; Step S200: Based on the total light tristimulus value, construct and solve an overdetermined system of equations to separate the tristimulus values ​​of each lamp light component and the tristimulus values ​​of the ambient light component. Specifically, this includes: Step S210: Based on the measured values ​​under each disturbance state, construct an overdetermined system of equations concerning the tristimulus values ​​of the lamp light component; Step S220: Solve the overdetermined system of equations using the weighted least squares method to obtain the optimal estimate of the tristimulus values ​​of the lamp light component; Step S230: Based on the baseline measured value and the optimal estimate of the tristimulus values ​​of the lamp light component, calculate the tristimulus value of the ambient light component. Step S300: Based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, calculate the compensation control command for the luminaire and control the luminaire to execute the corresponding compensation control command.

2. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 1, characterized in that, Step S100 includes: Step S110: Generate a corresponding multidimensional joint micro-perturbation sequence according to the preset calibration mode. The sequence includes a reference measurement step and multiple brightness micro-perturbation steps and color temperature micro-perturbation steps arranged in a preset order. Each step has preset perturbation parameters and preset duration. Step S120: Apply perturbations to at least one lamp in sequence according to the multidimensional joint micro-perturbation sequence; Step S130: In each perturbation step, the total light tristimulus value under the current state is synchronously acquired by the at least one optical sensor, and the average value is obtained by multiple samplings within the sampling window.

3. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 1, characterized in that, Step S220 includes: Step S221: Set a weight matrix according to the perturbation amplitude and signal-to-noise ratio of each perturbation step, wherein the measurement with the larger perturbation amplitude and the higher signal-to-noise ratio has a higher weight. Step S222: Based on the weight matrix, use the weighted least squares method to solve for the optimal estimate of the tristimulus values ​​of the light components of the lamp; Step S223: Calculate the residuals of each measurement value and the standard deviation of all residuals. Determine the corresponding measurement values ​​whose absolute residual values ​​exceed the preset multiple of the standard deviation as outliers and remove them. Reconstruct and solve the overdetermined system of equations based on the removed measurement values.

4. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 1, characterized in that, Following step S230, the following is also included: Step S240: For each sensor location, maintain a Kalman filter for real-time estimation and short-term prediction of the tristimulus values ​​of the ambient light component; the state vector of the Kalman filter at least contains the current tristimulus value estimate of the ambient light component at that sensor location and its rate of change. Step S250: Establish and maintain a daylight pattern model, record the tristimulus values ​​of ambient light components at each sensor location during different time periods each day, and learn the diurnal variation pattern of ambient light by performing exponential weighted moving average processing on multi-day data. Step S260: Using the daylight pattern model in conjunction with the Kalman filter includes: when the system starts up or is reinitialized, calling the daylight pattern model to provide an initial estimate of the ambient light component for the Kalman filter for the corresponding time period; during system operation, comparing the short-term prediction results of the Kalman filter with the long-term prediction reference values ​​of the daylight pattern model for the same time period, and if the deviation between the two exceeds a preset threshold, determining that the current illumination conditions are in an abnormal change state.

5. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 1, characterized in that, Step S300 includes: Step S310: Construct a light color transfer matrix based on the tristimulus values ​​of the light components of each lamp. The light color transfer matrix is ​​used to describe the influence of the output changes of the tristimulus values ​​of each lamp on the measured values ​​of the tristimulus values ​​at each sensor position. Step S320: Based on the light color transfer matrix, the tristimulus values ​​of the ambient light components, and the preset target lighting parameters, establish a quadratic programming optimization problem, where the decision variables are the output tristimulus values ​​of each lamp, the objective function is to minimize the total energy consumption of the system, and the constraints include physical model constraints on the tristimulus values ​​of each sensor location and lamp output range constraints. Step S330: Solve the quadratic programming optimization problem using a quadratic programming solver to obtain a multi-lamp collaborative control command that optimizes the objective function under the constraints, and control each lamp to execute the collaborative control command.

6. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 5, characterized in that, The process after step S330 also includes: Step S340: Use the S-curve function to generate a smooth transition curve for the lamp output from the current value to the target value. The expression is as follows: In the formula, The initial output value is for transition. The output value is the transition target, where t represents the current time starting from the transition start point. Let be the output value of the lamp at time t; The transition time is dynamically set according to the magnitude of output change; the greater the magnitude of change, the longer the transition time. This represents an S-shaped curve function; x represents the normalized time variable. Step S350: Set a control dead zone. When the absolute value of the deviation between the total light tristimulus value at the sensor position and the target illumination parameter is less than the preset sensing threshold, maintain the current output of the lamp and do not perform any adjustment. Step S360: Set the maximum allowable variation in brightness output and color temperature output of each lamp within a single control cycle to limit the output variation rate.

7. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 4, characterized in that, It also includes an adaptive calibration step, specifically including: real-time monitoring of the change rate of the ambient light component, selecting the corresponding calibration mode according to the interval in which the change rate is located, wherein the calibration modes are divided into a first calibration mode, a second calibration mode, and a third calibration mode in ascending order of calibration period; the first calibration mode, the second calibration mode, and the third calibration mode correspond to a multi-dimensional joint micro-perturbation sequence with an increasing number of execution steps, a more refined perturbation amplitude gradient, or a more comprehensive combination of perturbation types, respectively; the mode switching adopts a hysteresis mechanism, wherein the change rate trigger threshold for switching from a mode with a longer calibration period to a mode with a shorter calibration period is higher than the change rate trigger threshold for reverse switching, and the switching from a mode with a longer calibration period to a mode with a shorter calibration period requires the change rate to be lower than the corresponding degradation threshold and to last for a preset time.

8. The LED light field control method based on multidimensional micro-perturbation and spatial fusion as described in claim 7, characterized in that, The method is executed in a three-layer nested architecture comprising a first control layer, a second control layer, and a third control layer; the first control layer operates in a first control cycle and is used to perform long-term trend modeling of ambient light components, time-series prediction based on Kalman filters, and update learning tasks of daylight pattern models. The second control level operates with a second control cycle shorter than the first control cycle, and is used to perform multidimensional joint micro-perturbation sequence, optical component separation and optical color transfer matrix update tasks according to the mode determined by the adaptive calibration strategy; The third control level operates with a shorter control cycle than the second control cycle. Based on the output results of the first and second control levels, it performs the construction and solution of a quadratic programming optimization problem to generate collaborative control instructions. Finally, through smooth transition and anti-oscillation control steps, the instructions are sent to each lamp in real time.

9. A multi-dimensional micro-perturbation and spatial fusion LED light field control system, characterized in that, include: A multidimensional joint micro-perturbation control module is used to perform a multidimensional joint micro-perturbation sequence including brightness micro-perturbation and color temperature micro-perturbation on at least one luminaire in a lighting system, and the total light tristimulus value under each perturbation state is measured by at least one light sensor. The light component separation module is used to construct and solve an overdetermined system of equations based on the total light tristimulus values ​​to separate the tristimulus values ​​of the luminaire light component and the ambient light component. Specifically, it includes: Step S210: Constructing an overdetermined system of equations about the tristimulus values ​​of the luminaire light component based on the measured values ​​under each disturbance state; Step S220: Solving the overdetermined system of equations using the weighted least squares method to obtain the optimal estimate of the tristimulus values ​​of the luminaire light component; Step S230: Calculating the tristimulus values ​​of the ambient light component based on the baseline measured values ​​and the optimal estimate of the tristimulus values ​​of the luminaire light component. The adaptive compensation control module is used to calculate the compensation control command of the lamp based on the tristimulus values ​​of the separated ambient light components and the preset target lighting parameters, and to control the lamp to execute the corresponding compensation control command.

Images