CN-121510408-B - Closed loop LED driving system based on visible light sensor feedback

CN121510408BCN 121510408 BCN121510408 BCN 121510408BCN-121510408-B

Abstract

The invention discloses a closed-loop LED driving system based on visible light sensor feedback, which comprises a wireless communication module, a core control module, a voltage conversion module, a constant current driving module, a light source matrix module and a visible light sensor module. The invention collects the low-dimensional spectrum data of the LED light source in real time through the visible light sensor, reconstructs the low-dimensional spectrum data into high-dimensional spectrum power distribution by utilizing the neural network model on the microprocessor, dynamically adjusts the driving current of the multichannel LED by PID closed-loop control based on the difference between the reconstructed spectrum and the target spectrum, realizes accurate tracking and correction, effectively overcomes the influence of factors such as aging of the LED device, environmental change and the like, ensures the stability and consistency of the output spectrum, and improves the adaptability and the intelligent degree of the light source.

Inventors

HUANG WEIRUN
HUANG PENGJU
YUAN HUAN
Zeng Runqi
HE JIAQI
HONG YINGHAN
CAO HUI
MENG LIANGYU
CHEN YURONG
WANG JINBO
ZHU ZEKUN
WANG ZHE
OU YI
WEN CHUNLEI

Assignees

广州航海学院

Dates

Publication Date: 20260505
Application Date: 20260113

Claims (7)

1. A closed loop LED driving system based on visible light sensor feedback comprises a wireless communication module, a core control module, a voltage conversion module, a constant current driving module, a light source matrix module and a visible light sensor module, and is characterized in that, The wireless communication module is connected with the core control module, the output end of the core control module is connected with the input end of the constant current driving module, the output end of the voltage conversion module is connected with the input end of the constant current driving module, the output end of the constant current driving module is connected with the input end of the light source matrix module, the output end of the light source matrix module is connected with the input end of the visible light sensor module, and the output end of the visible light sensor module is connected with the core control module; The core control module comprises a microprocessor, wherein the microprocessor comprises a spectrum reconstruction unit and a closed loop feedback control unit; the microprocessor performs real-time spectrum reconstruction through the spectrum reconstruction unit and performs closed-loop feedback through the closed-loop feedback control unit; the real-time spectrum reconstruction specifically comprises the following steps: s1-1, collecting spectrum data of different light sources, generating a training sample, preprocessing the data in the training sample, generating a preprocessed training sample, and dividing a training set, a verification set and a test set; s1-2, constructing a neural network model, wherein the neural network model is configured to reconstruct low-dimensional spectral power distribution data into high-dimensional spectral power distribution data; S1-3, optimizing hidden layer neurons and learning rate through Optuna based on the training set, and training on the neural network model by combining Adam, earlyStopping and ModelCheckpoint; S1-4, evaluating the spectrum reconstruction performance of the model based on the verification set and the test set, and selecting an optimal model; S1-5, converting the optimal model into TFLite and deploying the TFLite into an embedded system, outputting the high-dimensional spectral power distribution data by the microprocessor based on the input data of the visible light sensor, and calculating a correlated color temperature CCT, a color rendering index CRI and color coordinates to obtain a reconstructed spectrum; the closed loop feedback is specifically: s2-1, receiving a target correlated color temperature set by a user, and acquiring a corresponding target spectrum; S2-2, comparing the reconstructed spectrum with the target spectrum, and calculating a spectrum error and a colorimetry error; s2-3, generating adjustment amounts of LEDs of all channels through a PID closed loop system according to the error signals; s2-4, updating PWM driving signals of all channels according to the adjustment quantity of the LEDs of all channels, and adjusting the luminous intensity of the LEDs to correct spectrum deviation; S2-5, continuously collecting, reconstructing, calculating and adjusting until the error between the output spectrum and the target spectrum is within a preset threshold value.
2. The closed-loop LED driving system based on feedback of a visible light sensor according to claim 1, wherein in S2-2, the chromaticity error is calculated by calculating tristimulus values according to a spectral power distribution SPD, calculating normalized chromaticity coordinates according to the tristimulus values, and finally obtaining the chromaticity error, wherein the tristimulus values are calculated by: ; ; ; Wherein k is a normalization coefficient; as a spectral power distribution function; 、 And Is CIE 1931 2 DEG standard color matching function; is the sampling wavelength.
3. The closed-loop LED driving system based on feedback of visible light sensor according to claim 1, wherein in S2-4, the updating PWM driving signals of each channel comprises the following specific calculation processes: ; ; Wherein, the light mixing matrix Is a2 x 9 underdetermined matrix; is a light mixing matrix An inverse matrix of (a); An output matrix for the PID closed loop system; The adjustment amount of the LEDs of each channel is adjusted; the latest duty cycle data for the ith channel; duty cycle data last time for the ith channel; LED adjustment for the ith channel, wherein 。
4. A closed loop LED driving system based on feedback of a visible light sensor according to claim 1, wherein in S2-3, the output of the PID controller in the PID closed loop system is determined by a proportional term P, an integral term I and a differential term D together: Wherein, the Is a control amount; is a proportionality coefficient; Is an integral coefficient; Is a differential coefficient; is a colorimetry error signal; a function of integrating the colorimetry error signal.
5. The closed-loop LED driving system based on visible light sensor feedback according to claim 1, wherein the wireless communication module comprises a WiFi module and a mobile terminal application program APP, the mobile terminal application program APP is used for sending a control instruction to the WiFi module through a wireless communication protocol according to the target correlated color temperature preset by a user, and the WiFi module is used for analyzing received data and then transmitting the data to the core control module.
6. The closed loop LED driving system based on feedback of visible light sensor according to claim 1, wherein the light source matrix module is composed of a combination of seven LED chips of different wavelengths.
7. The closed loop LED drive system based on visible light sensor feedback of claim 1, wherein the visible light sensor module comprises a multi-channel visible light sensor having an independent channel defined spectral response.

Description

Closed loop LED driving system based on visible light sensor feedback Technical Field The invention relates to the technical field of LED illumination, in particular to a closed-loop LED driving system based on visible light sensor feedback. Background To overcome the limitations of open loop systems, the introduction of closed loop spectral control is an important technical direction. In recent years, with the development of visible light sensor technology, some high-integration multi-channel visible light sensors, such as micro spectrometers based on filter arrays or diffraction gratings, have emerged, which can provide spectral data of several tens of wavelength points in the 380-780nm range. Such sensors offer new possibilities for closed loop spectral control. By introducing real-time spectrum feedback, the system can acquire a more complete output spectrum, and then by combining an optical model, colorimetry calculation and an optimization algorithm, the driving parameters of the LEDs of each channel can be regulated more accurately, so that dynamic tracking and correction of a target spectrum are realized. In the prior art, chinese patent document CN119946951a discloses an AI plant light spectrum adjusting method for optimizing plant photosynthesis, which collects plant physiological data through an intelligent spectrum sensor network, constructs a three-dimensional spectrum response mathematical model, and obtains a plant spectrum matching network by knowledge distillation training, and combines plant growth stage evaluation and differential co-evolution algorithm to generate an LED dimming scheme, so as to realize personalized spectrum optimization for different plant individuals and growth stages. However, the scheme is mainly applied to the field of agricultural plant illumination, and aims to improve photosynthetic efficiency, reduce energy consumption and give consideration to economic cost, emphasize suitability for plant growth stage and physiological characteristics, and cannot meet the requirements of light quality stability and color rendering in indoor and outdoor illumination environments. In the prior art, chinese patent document CN120239144a discloses an intelligent lighting control system and method based on a multi-mode sensor, which collects multi-dimensional data such as environmental optics, thermal, acoustic, spectrum, ultraviolet and human motion through a multi-mode sensor network, combines with feature calculation, space-time coupling modeling, multi-mode verification and biorhythm compensation modules, and finally realizes multi-dimensional dynamic regulation and control of lighting parameters through LSTM-transducer hybrid architecture and rolling time domain optimization strategy. However, the scheme focuses on intelligent architecture or medical places, emphasizes photo-thermal coupling analysis, ultraviolet exposure safety control and biorhythm compensation to ensure environment suitability and human health safety, and is not suitable for general illumination scenes with high requirements on light quality. Therefore, it is desirable to provide a closed-loop LED driving system based on feedback of a visible light sensor, which is oriented to a general illumination scene, so as to solve the stability problem of the output spectrum in the prior art. Disclosure of Invention The invention provides a closed-loop LED driving system based on visible light sensor feedback, which solves the problem that an open-loop system in the prior art is difficult to ensure long-term stable spectrum consistency. The system realizes high-precision, low-delay and strong-robustness spectrum control capability through the cooperative optimization of the neural network and the control strategy, provides a feasible path for improving the light quality of the intelligent lighting system, and has application potential for being further popularized to medical lighting, museum lighting, precision agriculture and other scenes. The technical scheme of the invention is realized by that the closed-loop LED driving system based on visible light sensor feedback comprises a wireless communication module, a core control module, a voltage conversion module, a constant current driving module, a light source matrix module and a visible light sensor module; The wireless communication module is connected with the core control module, the output end of the core control module is connected with the input end of the constant current driving module, the output end of the voltage conversion module is connected with the input end of the constant current driving module, the output end of the constant current driving module is connected with the input end of the light source matrix module, the output end of the light source matrix module is connected with the input end of the visible light sensor module, and the output end of the visible light sensor module is connected with the core control module; The core control module