Secure Transmission in Cellular V2X Communications Using Deep Q-Learning

Abstract

Cellular vehicle-to-everything (V2X) communication is emerging as a feasible and cost-effective solution to support applications such as vehicle platooning, blind spot detection, parking assistance, and traffic management. To support these features, an increasing number of sensors are being deployed along the road in the form of roadside objects. However, despite the hype surrounding cellular V2X networks, the practical realization of such networks is still hampered by under-developed physical security solutions. To solve the issue of wireless link security, we propose a deep Q-learning-based strategy to secure V2X links. Since one of the main responsibilities of the base station (BS) is to manage interference in the network, the link security is ensured without compromising the interference level in the network. The formulated problem considers both the power and interference constraints while maximizing the secrecy rate of the vehicles. Subsequently, we develop the reward function of the deep Q-learning network for performing efficient power allocation. The simulation results obtained demonstrate the effectiveness of our proposed learning approach. The results provided here will provide a strong basis for future research efforts in the domain of vehicular communications.