CIR fast approximation for closed space VLC transmission with the use of a trained artificial neural network

Abstract

The channel impulse response (CIR) fast approximation approach for the visible light communication (VLC) in closed spaces based on a previously trained model of artificial neural networks of deep learning (ANN-DL) was presented. There were used during the training sets the CIRs characteristics, obtained on a way of numerical simulations based on modified Monte Carlo ray tracing algorithm (MMC). The Matlab Deep Learning Toolbox software library was applied as the environment to the model preparation.