Abstract
This paper proposes a controlled signal technique for visible light non-orthogonal multiple access (VL-NOMA) communication in an interference-controlled environment with intelligent reflecting surfaces (IRS) for beyond 5G (B5G) and 6G communication networks. The light-emitting diode (LED) is used for carrier signal generation to transmit signals to the two users (photodiodes, PDs) due to its advantages, such as its programmable nature and flexibility. The potential challenge is how the signals could be controlled with an IRS approach, which prompted this research. We have used IRS, which is a cutting-edge enabling technology that modifies the signal's reflection by utilizing numerous inexpensive passive reflecting elements to improve the signal's performance. Furthermore, deep reinforcement learning (DRL) is deployed to control the reflected signals, simulate, make decisions, and link LED-IRS-PDs, redirecting the signals. The entire system is successfully synchronized, and then the bit error rate (BER), line of sight (LOS), and non-line of sight (NLOS) performances are investigated. Furthermore, we place a blocker at the center of the model as a NLOS to check how the transmitted signals will perform. We observed that the propagated signal improved the BER as per LOS, hence, the NLOS blocker reduced the signal's performance. Furthermore, we optimized the signals to investigate BER, LOS, and NLOS signal performance. We observed that LOS signals performed better than NLOS signals.