The performance improvement of visible light communication systems under strong nonlinearities based on Gaussian mixture model

Nonlinearity is a major problem in visible light communication (VLC). In recent years, machine learning has emerged and has been widely used in many fields. Among them, the nonlinear algorithms have gradually become a powerful tool to solve the nonlinearity in VLC. In this article, we compare the performance of two main clustering algorithms—Gaussian mixture model (GMM) and K‐means under strong nonlinear conditions. Experimental results show that the signal peak‐to‐peak voltage (Vpp)r range is approximately 0.25 V that meets the BER threshold of forward error correction when using GMM and is larger than the range using K‐means under the data rate is 1.5 Gbps.