Efficient hybrid signal detection for optical NOMA with 512-QAM in visible light communication

This work presents a hybrid signal detection scheme for optical non-orthogonal multiple access (NOMA) systems using 512-QAM modulation in visible light communication (VLC) networks. As future VLC systems demand ultra-high data rates and superior spectral efficiency, they require high-order modulation like 512-QAM, which is highly susceptible to noise, nonlinearities, and channel distortions, complicating reliable detection. To tackle these challenges, a hybrid detection strategy combining traditional algorithms – such as minimum mean square error (MMSE) and zero-forcing (ZF) – with advanced machine learning techniques like deep neural networks (DNNs) is proposed. This approach harnesses the low complexity and speed of conventional detectors alongside the adaptive learning and nonlinear handling capabilities of DNNs, enabling robust detection under challenging conditions. Simulations conducted under Rayleigh fading in a MIMO-VLC setup using MATLAB 2016 extensively evaluate BER and PSD performance. Results demonstrate that the hybrid model achieves the target BER of 10 −3 with up to a 13 dB SNR improvement over traditional methods. Moreover, it ensures better spectral containment, minimizing interference and enhancing bandwidth utilization. These outcomes highlight the potential of hybrid detection schemes to realize high-capacity, energy-efficient, and reliable VLC systems, making them suitable for smart indoor environments such as wireless data hubs, augmented reality, and smart lighting in hospitals and intelligent buildings.