Machine Learning in Visible Light Communication System: A Survey

Table 2

ML algorithms in channel estimation.

Method	Ref.	Outcomes
ANN	[10]	(1) Simple to use; (2) Precision level of up to 97.7% can be obtained
Adaptive PBL	[12]	(1) Can estimate real-time indoor VLC channel; (2) Require less training time; (3) Decrease calculation complexity
Bayesian compressive sensing	[13]	(1) It can be used to predict underwater VLC channels; (2) Enhance efficiency; (3) Increase prediction correctness
DNN	[11]	(1) Improve BER without requiring any complicated calculations; (2) It has BER performance superior to LS and MMSE algorithms