Reducing PAPR in optical NOMA-based visible light communication systems for beyond 5G applications

This paper investigates the reduction of Peak-to-Average Power Ratio (PAPR) in Optical Non-Orthogonal Multiple Access (O-NOMA) based Visible Light Communication (VLC) systems for beyond fifth generation (B5G) wireless applications. High PAPR is a major challenge in O-NOMA systems because it increases signal distortion and reduces the efficiency of optical transmitters. To address this issue, a deep learning assisted optimization technique based on Recurrent Neural Network–Partial Transmit Sequence (RNN-PTS) is proposed. The performance of the proposed approach is evaluated using Complementary Cumulative Distribution Function (CCDF) analysis for 256, 512, and 1,024 sub-carriers and compared with conventional techniques such as Clipping, Selective Mapping (SLM), Partial Transmit Sequence (PTS), Genetic Algorithm-PTS (GA-PTS), and Particle Swarm Optimization-PTS (PTS-PSO). Simulation results show that the proposed RNN-PTS method significantly reduces PAPR, achieving approximately 4.5 dB, 5.8 dB, and 6.0 dB at a CCDF of 10 −3 for 256, 512, and 1,024 sub-carriers, respectively. Furthermore, Bit Error Rate (BER) analysis demonstrates improved communication reliability with lower Signal-to-Noise Ratio (SNR) requirements. Power Spectral Density (PSD) analysis also confirms better spectral containment and reduced out-of-band radiation. Overall, the proposed RNN-PTS approach improves power efficiency, spectral efficiency, and transmission reliability in O-NOMA based VLC systems.