An Adaptive Network Selection Method for Enhancing Physical Layer Security in Heterogeneous Vehicular Networks

The heterogeneous vehicular networks (HVNs) integrate various radio access technologies to enhance user capacity, coverage, and quality of experience. However, the rapid topology changes in HVNs make it challenging for traditional encryption methods that rely on trusted third parties (TTPs) to store, manage, and distribute keys. Physical layer security (PLS) technology emerges as a powerful complementary solution due to its capabilities in key generation and distribution without relying on TTPs. Therefore, we derive two PLS metrics for HVNs, connection outage probability (COP) and secrecy outage probability (SOP). Based on these metrics, we propose an adaptive network selection method which enhances PLS by identifying and selecting the network that is optimal secure and reliable. Specifically, to characterize the irregularity and randomness of HVNs, Poisson point processes and Poisson line processes are adopted to build a spatial model for HVNs. Then, we derive closed-form expressions for COP and SOP in HVNs, which characterize the connectivity stability and security of networks. Furthermore, we apply fuzzy logic to filter available networks, aiming at reducing computational latency. Analytic hierarchy process and entropy weight method are utilized to quantitatively analyze the diverse requirements of in-vehicle services and network metrics. Finally, the technique for order of preference by similarity to ideal solution is used to accurately select the optimal network from the filtered candidates, enhancing the stability and security of vehicle communication in HVNs. Simulation results demonstrate that our proposed method significantly enhances the PLS of HVNs, achieving a maximum of roughly 22.29% improvement in SOP and an average increase of 29.58% in COP performance compared to conventional methods.