Optimizing intelligent reflecting surface assisted visible light communication networks under blockage and practical...

Wireless communication systems can enhance their capabilities by exploring new opportunities and addressing emerging challenges through the integration of the Internet of Things (IoT) in 6G networks. Visible Light Communication (VLC) stands out as a promising wireless access technology for IoT devices. This paper presents a novel Teaching-Learning-Based Optimization (TLBO) optimized Intelligent Reflecting Surface (IRS)-assisted VLC system aimed at maximizing Signal-to-Noise Ratio (SNR) and enhancing illuminance uniformity. The proposed system ensures improved communication for IoT devices, particularly in indoor environments. To enhance real-world applicability, the study evaluates system performance under realistic conditions, considering the impact of human and furniture blockages. A detailed investigation is conducted on the achievable SNR performance of the TLBO-optimized IRS-VLC system, demonstrating its ability to maintain robust and reliable communication in obstructed areas. Furthermore, the impact of the proposed system on illuminance uniformity is thoroughly analyzed, showing significant improvements in lighting efficiency and visibility. The effectiveness of the approach is validated using the Coefficient of Variation (CoV), confirming that the system achieves an optimal CoV range of 0-0.2 across the room, ensuring uniform illumination for IoT applications. The SNR and received power improve with increasing IRS elements, with optimal device positioning (e.g., edge-mid) outperforming suboptimal configurations (e.g., corner-opposite) due to geometric reflection efficiency and aperture gain, leading to a widening performance gap at larger IRS scales-highlighting the critical role of strategic IoT device placement in fixed sensor deployments. The findings highlight the potential of TLBO-optimized IRS-VLC systems in overcoming physical obstructions and improving both communication and lighting conditions in smart indoor environments. Additionally, detailed convergence analysis demonstrates that TLBO performs better than Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) in terms of convergence speed, higher fitness value and lower sensitivity to initial conditions, making it most suitable for real-time IRS-VLC based IoT applications.