10 Gbps RoFSO-VLC hybrid system design and performance analysis with machine learning
PHOTONIC NETWORK COMMUNICATIONS, cilt.52, sa.1-2, 2026 (SCI-Expanded, Scopus)
- Yayın Türü: Makale / Tam Makale
- Cilt numarası: 52 Sayı: 1-2
- Basım Tarihi: 2026
- Doi Numarası: 10.1007/s11107-026-01055-0
- Dergi Adı: PHOTONIC NETWORK COMMUNICATIONS
- Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Aerospace Database, Applied Science & Technology Source, Compendex, INSPEC, Materials Science & Engineering Collection (ProQuest), Technology Collection (ProQuest)
- Gazi Üniversitesi Adresli: Evet
Özet
In this study, the design of Radio-over-Free Space Optic-Visible Light Communication (RoFSO-VLC) hybrid system is implemented using Optisystem software for optical simulation. For the optical performance evaluation, the Q-factor and Bit Error Rate (BER) data were determined using various parameters that affect the overall system performance (laser power, FSO attenuation, FSO link distance, VLC link distance, and VLC frequency values for different atmospheric conditions). The evaluation was based on the criteria that the BER should be no higher than 10(-)(9) and the Q-factor of at least 6 was considered for efficient transmission. Also in this study, the quality factor (Q-factor) of hybrid system is estimated by using different machine learning algorithm (MLA) models such as linear support vector machine (SVM), Gaussian progress regression, fine tree regression, narrow neural networks, and artificial neural networks (ANN) using MATLAB (R) program. The data obtained with the simulation program are used at train and test phases of these MLA models, and the best prediction performance is tried to be obtained. This study aims to create a hybrid structure for more efficient transmission of RF signals by modulating them on the FSO link and transferring them to the end user via the VLC system as an indoor connection and, at the same time, to obtain the prediction model of this system using machine learning (ML) methods and to propose a method for creating ML models for the prediction of system performance. As a result, models that can predict around 90% and above were obtained in 67% of the prediction studies conducted with 5 different ML algorithms and models, and the best performance of prediction was provided by the ANN model (98%).