Hybrid RF-VLC Communication System: Performance Analysis and Deep Learning Detection in the Presence of Blockages

This paper analyzes the performance of a dual-hop hybrid radio frequency - visible light communications (VLCs) system. VLC exploits high frequencies for signal propagation for short-range and is highly susceptible to blockages. Hence, in this work, the sensitivity of the VLC toward the blockages is explored. This study examines outdoor-indoor communication for an urban scenario featuring an office environment. The VLC system is designed by considering multiple VLC access points serving the end users in the presence of human blockages. Selection combining is employed to select the access point which offers the maximum instantaneous signal-to-noise ratio at the user. To capture the outdoor scenario, path loss modeling is conducted to account for signal attenuation from outdoor to indoor spaces. The outdoor scenario is modeled using Nakagami-m fading channels while VLC is employed for the indoor scenario. System performance is assessed in closed-form expression for outage probability and the symbol error rate for binary phase shift keying and quadrature amplitude modulation. A deep learning-based approach detects the symbols by tracking dynamic changes in the channel. Simulation results corroborate the correctness of derived analytical expressions and reveal that blockages significantly impact both single and multiple LED based communication channels.