| Summary: | 碩士 === 國立臺灣科技大學 === 自動化及控制研究所 === 100 === The advancement in semiconductor materials and processing technologies has improved the color performance and brightness of light-emitting diode (LED). The spectral distribution of LED is quite narrow; therefore it requires various wavelengths of LEDs and adjusting spectral intensity value of the LEDs to obtain the simulated spectrum.
But a mathematically solvable process with many variables is not only complex but also time consuming. Manual calculation is not easy to obtain the optimal values and has the risk of pilot errors. Compared with that, the Genetic Algorithm is an efficient way to acquire for the optimal values of the parameters for spectral intensity value of the LEDs.
This research uses 5 different types of target spectra in the following order: D65, A, Gaussian spectral intensity, equal spectral intensity, and square,within the spectral range from wavelength 300 nm to 700 nm. The simulate spectra were composed of adopting 7 to 11 different LED wavelengths. The values of the Root-Mean-Square (RMS) error for these 5 spectrum types are 4.91 %, 1.93 %, 2.07 %, 2.80 %, and 12.30 %, respectively.
A spectrometer program and a control circuit is developed. It first runs the Genetic Algorithm to acquire the parameters, which are then passed to control the intensities of the LEDs. By using a spectrometer, the spectrum of the LED is passed to the computer program to process the feedback control.
We have demonstrated a feedback control system to simulate the standard illuminants by a combination of LEDs.
|
|---|
