De-Multiplexing of NOMA VLC Signals Using Probabilistic Neural Network

Visible light communications (VLC) and non-orthogonal multiple access (NOMA) are promising technology expected to address key challenges in the next generation of wireless networks. However, optical wireless communications in indoor environments have several unique characteristics, such as slow fading and similar channel conditions. In such circumstances, the traditional successive interference cancellation (SIC) de-multiplexing method for NOMA cannot perform well. In this paper, for the first time, we proposed a de-multiplexing process using a probabilistic neural network (PNN). We presented in details the implementation process and experimental demonstration of a NOMA VLC system. Experimental results indicated that the PNN method was more robust against interference than the SIC method.