Service Route Optimization for Autonomous Vehicles in Edge-Assisted Cellular Infrastructure

초록

In this paper, we address the challenge of optimizing service routes for autonomous vehicles operating within edge-assisted cellular infrastructures. To tackle the complexity of resource-constrained and dynamic environments, we introduce a three-stage service routing model that spans uplink base stations, edge servers, and downlink base stations. We first prune infeasible links based on real-time resource availability and then logically expand the network into multi-layered structures to streamline route optimization. Building on this, we propose the SErvice Route Optimization (SERO) algorithm, which leverages Lagrangian dual decomposition to allocate resource-efficient service routes while preventing computational and network bottlenecks. Extensive simulations based on real-world city maps and vehicular traces from 2,502 vehicles demonstrate that SERO significantly reduces service failure rates-by up to 68.27% compared to existing approaches. Our findings underscore the importance of jointly optimizing communication and computation resources for scalable and reliable autonomous vehicle services in future B5G/6G networks. © 2025 IEEE.

키워드