Enhancing VLC Vehicle Networks in Channel Estimation, Coding, and Multiple Access

Vehicle communications’ low latency requirement and high moving velocity prevent traditional technologies in wireless communications to work well in Visible Light Communication (VLC). This study presents a VLC system with enhanced channel estimation accuracy, coding performance, and data rate. To estimate the highly fluctuating vehicular channel, we propose a neural network-based channel estimation method that leverages non-pilot symbols, achieving over 1dB improvement compared to baselines in simulation. Additionally, Adjacent-Bits-Swapped+ (ABS+) polar codes with short lengths and low rates achieve a 0.7dB-2.5dB gain and reduce decoding time to one-quarter of standard Polar list decoding, without affecting the bit error rate. Furthermore, a three-stage power and subcarrier allocation scheme for Sparse Code Multiple Access (SCMA) is introduced to improve overall throughput in multi-vehicle scenario by over 1dB.