AL-aided AMC in a multi-user white-light OFDMA VLC system over a light-diffusing fiber loop.

Area of Science:

• Optical Communications
• Artificial Intelligence in Communications
• Wireless Networking

Background:

• Visible Light Communication (VLC) systems offer high-speed wireless data transmission using illumination.
• Adaptive modulation is crucial for optimizing performance in dynamic multi-user environments.
• Orthogonal Frequency Division Multiplexing Access (OFDMA) enables efficient spectrum utilization in VLC.

Purpose of the Study:

• To develop an adaptive modulation scheme for flexible, high-speed, multi-user VLC.
• To implement Automatic Modulation Classification (AMC) for monitoring modulation formats in OFDMA-based VLC.
• To enhance the efficiency and accuracy of modulation classification in VLC systems.

Main Methods:

• A joint Convolutional Neural Network (CNN), Active Learning (AL), and Data Augmentation (DA) approach for AMC.
• Integration of a diffuse white-light VLC system with an OFDMA framework.
• Experimental demonstration with a two-user access scenario and unequal bandwidth allocation.

Main Results:

• Achieved data rates of 325.5 Mbps with a Bit Error Rate (BER) of 2.163 × 10^-3 within the Forward Error Correction (FEC) threshold.
• Demonstrated a classification accuracy rate of 95.48% for OFDMA signal subcarriers using the AL-aided CNN AMC scheme.
• Showcased faster convergence compared to traditional CNN-based AMC with limited training data (240 samples).

Conclusions:

• The proposed AL-aided CNN AMC scheme effectively supports adaptive modulation for high-speed, multi-user VLC.
• The system provides reliable white-light illumination and ubiquitous access.
• This approach enhances spectral efficiency and adaptability in next-generation VLC networks.