Sum Rate Maximisation for IRS-Assisted VLC Using Reinforcement Learning

The increasing demand for high-speed and reliable communication systems has highlighted the potential of visible light communication (VLC), particularly for indoor wireless connectivity. However, VLC systems face challenges due to probabilistic factors affecting the line-of-sight (LoS) link quality, which is crucial for efficient data transmission. This paper addresses these challenges by applying reinforcement learning (RL) with linear function approximation to maximise the sum rate of VLC systems assisted by intelligent reflecting surfaces (IRSs). The proposed algorithm optimises the allocation of the IRS elements to the system users, effectively mitigating the impact of link blockages and random device orientation. Simulation results demonstrate the efficacy of the reinforcement learning approach in improving the sum rate and resolving blockage issues.