A Reliable Zero-Trust Network for Task Offloading in Vehicular Systems Using an Asynchronous Federated Learning Approach in 6G

Abstract

In the emerging 6G era, vehicles are extensively connected to wireless networks through edge-accessible roadside units (RSUs). The increasing number of connected vehicles and vehicle services introduces a significant security challenge known as the "zero-trust network (ZTN)." This necessitates a shift from traditional methods of resource slicing and scheduling. This study focuses on ensuring reliable 6G vehicular services, particularly addressing the scenario of task offloading between vehicles, which involves managing communication resources. We propose a method that uses a logical model to assign an edge node score (ENS) to evaluate the security of edge nodes, thereby protecting vehicles from potential threats posed by untrusted edge access points. Vehicles select edge nodes with high ENS scores for task offloading. Also, we used a federated asynchronous reinforcement learning approach to enhance the management of offloaded tasks. Simulation results show that the proposed approach effectively organizes the resources and ensures the security of vehicle data.