Visible light communication for vehicle-to-vehicle systems: a deep neural network-based signal detection framework

Abstract Visible light communication (VLC) is emerging as a promising alternative to conventional radio frequency-based vehicle-to-vehicle (V2V) communication due to its high data rate, immunity to electromagnetic interference, and use of energy-efficient LED-based vehicular lighting systems. However, signal detection in dynamic V2V-VLC environments remains a significant challenge due to factors such as high mobility, channel fading, ambient light interference, and misalignment between transmitter and receiver. This paper proposes a robust deep neural network (DNN) architecture specifically designed for accurate and reliable signal detection in V2V-VLC systems. Through extensive simulations and experimental evaluations under varying vehicular conditions, our framework achieves superior detection accuracy of 99.8 % at a distance of 1 m and 98.4 % at 5 m while exhibiting greater resilience compared to traditional signal processing methods. The research aims to enhance communication reliability in intelligent transportation systems, paving the way for safer and more efficient autonomous driving environments.