Challenges and prospects of machine learning in visible light communication

Visible light communication (VLC) is a promising research field in modern wireless communication. VLC has its irreplaceable strength including rich spectrum resources, no electromagnetic disturbance, and high-security guarantee. However, VLC systems suffer from the non-linear effects that exist in almost every part of the system. As a part of artificial intelligence, machine learning (ML) is showing its potential in non-linear mitigating for its natural ability to fit all kinds of transfer functions, which may dramatically push the research in VLC. This paper introduces the application of ML in VLC, describes five recent research of deep learning applications in VLC, and analyses the performance.