Underwater optical wireless communication system: Deep learning CNN with NOMA-based performance analysis

Abstract This research is looking forward improving the performance for underwater optical wireless communication (UOWC) by applying a Non-orthogonal multiple access (NOMA) technique. We also get benefit of advantage transmission based on convolutional neural network hybrid with long short-term memory cell. The relays selection and power optimization are two main parameters to enhance UOWC system performance. In this work, we suppose pairing method NOMA nodes. By replacing inner dense connections convolution layers, model proposed overcome high complexity over fitting improve obtained sum rates show that outperforms orthogonal ~ 6%. Applying step-by-step sub-optimization algorithm (SSOPA) yields better results than using fixed allocation (FPA), while global optimal (GOPA) increases both FPA SSOPA. It found improvement when GOPA combined CNN approach enhances 2.5% independent-relay-aided (ICNOMA) UOWC. 1.2%, 2.5%, 8.7% 0.12%, 0.34%, 2.09% SSOPA, clear, pure, coastal water, respectively. ICNOMA ordinary (ONOMA) cooperative (CNOMA) without independent relay achieves an ONOMA CNOMA 20.4% 3.2%,