Capacity analysis of infrastructure-to-vehicle visible light communication with an optimized non-orthogonal multiple access scheme

Abstract

Non-Orthogonal Multiple Access (NOMA) is considered an effective solution to significantly improve the performance of Visible Light Communication (VLC) networks. In this article, we investigate Bit-Error Rate (BER), capacity, and fairness of the NOMA-based infrastructure-to-vehicle (I2V) VLC system. First, toward realistic channel modeling, we consider the 3D CAD model of the I2V VLC system and the commercial radiation pattern of the red traffic light that acts as a wireless transmitter. The impact of both longitudinal and lateral shifts between traffic light and vehicles is also considered. Then, based on the adopted realistic channel path loss model and using NOMA, we investigate BER and capacity performances, assuming different possible I2V scenarios. Moreover, we obtain the optimal power allocation coefficient that maximizes the sum capacity and the fairness index. The problem under investigation is first formulated, and then the optimal solution is obtained using the Exhaustive Search (ES) method. However, to reduce the time complexity, we further propose different sub-optimal solutions, which depend on Genetic Algorithms (GAs), Particle Swarm (PS), and Gray Wolf Optimization (GWO) methods. The obtained results demonstrate that the adopted sub-optimal solutions provide very close values to the optimal solution. The results further indicate a significant performance improvement due to utilization of the deduced optimized coefficient. Considering Jain’s fairness, utilization of the obtained optimal power allocation coefficient achieves a very high fairness index (≈ 1), while satisfying a target BER of less than 10 - 3 for all I2V scenarios under consideration.