A Learning Approach for Robust Carrier Recovery in Heavily Noisy Visible Light Communication

Abstract : Visible Light Communication (VLC) exploits optical frequencies, diffused by usual LED lamps, for adding data communication features to illuminating systems. This paradigm has attracted a growing interest in both scientific and industrial community in the latter decade. Nevertheless, classical wireless communication mechanisms for physical and Medium Access Control (MAC) layers are hardly available for VLC, due to the massive external interference caused by sunlight. A correct signal carrier recover in high noise conditions represent a significant challenge. In this work, it is shown that the synchronization frame length affects the performance of the system in terms of Bit Error Ratio (BER). Since different external conditions require different minimum preamble lengths, we considered an Artificial Intelligence (AI) approach, based on multi-arm bandit formulation, for obtaining a low impact in both BER and goodput of the communication. A low-cost hardware VLC system, implementing a learning algorithm on a Frequency Shift Keying Modulation (FSK), has been designed and tested in different environmental conditions. Experimental results show that a proper choice of preamble length overcomes, in terms of BER and goodput, the classical approach based on fixed preambles.
Complete list of metadatas

Cited literature [2 references]  Display  Hide  Download

Contributor : Antonio Costanzo <>
Submitted on : Monday, February 18, 2019 - 10:09:29 AM
Last modification on : Thursday, February 21, 2019 - 4:52:32 PM
Long-term archiving on : Sunday, May 19, 2019 - 3:02:13 PM


Files produced by the author(s)


  • HAL Id : hal-02022625, version 1



Antonio Costanzo, Valeria Loscri. A Learning Approach for Robust Carrier Recovery in Heavily Noisy Visible Light Communication. WCNC 2019 - IEEE Wireless Communications and Networking Conference, Apr 2019, Marrakesh, Morocco. ⟨hal-02022625⟩



Record views


Files downloads