Machine Learning Enabled LED Lighting Using Scattering Optics

Abstract Illumination receives a great deal of attention as white light‐emitting diodes become energy‐efficient light sources in households and commercial buildings, on streets and highways, and at stadiums and construction sites. In general, lenses and mirrors are used to control the spatial distribution of white LED (WLED) light. Here, it is proposed to use an optical diffuser, the key device in scattering optics, to achieve a pre‐defined WLED brightness distribution by nanocrystals and machine learning. Optical diffusers are typically used to create soft light (similar brightness from any angle of view), however, here the concentration of nanocrystals in a nanocomposite film (optical diffuser) to tune its optical property at different regions is altered. Machine learning is employed to achieve the inverse design of the optical diffuser pattern controlling the WLED brightness distribution, and this design task is beyond human capacities which are carried out using the brute force approach. In the end, several pre‐defined WLED brightness distributions are demonstrated for showing the success of this efforts.