Energy Efficient Orchestration in Multiple-Access Vehicular Aerial-Terrestrial 6G Networks

Abstract

The proliferation of users, devices, and novel vehicular applications - propelled by advancements in autonomous systems and connected technologies - is precipitating an unprecedented surge in novel services. These emerging services require substantial bandwidth allocation, adherence to stringent Quality of Service (QoS) parameters, and energy-efficient implementations, particularly within highly dynamic vehicular environments. The complexity of these requirements necessitates a fundamental paradigm shift in service orchestration methodologies to facilitate seamless and robust service delivery. This paper addresses this challenge by presenting a novel framework for service orchestration in Unmanned Aerial Vehicles (UAV)-assisted 6G aerial-terrestrial networks. The proposed framework synergistically integrates UAV trajectory planning, Multiple-Access Control (MAC), and service placement to facilitate energy-efficient service coverage while maintaining ultra-low latency communication for vehicular user service requests. We first present a non-linear programming model that formulates the optimization problem. Next, to address the problem, we employ a Hierarchical Deep Reinforcement Learning (HDRL) algorithm that dynamically predicts service requests, user mobility, and channel conditions, addressing the challenges of interference, resource scarcity, and mobility in heterogeneous networks. Simulation results demonstrate that the proposed framework outperforms state-of-the-art solutions in request acceptance, energy efficiency, and latency minimization, showcasing its potential to support the high demands of next-generation vehicular networks.

Figures (8)

• Figure 1: System Model: Supporting UEs through RSUs and UAVs with service coverage, quality channel access, and deployed function availability.
• Figure 2: The proposed method's process for time frame $t$, including Information Gathering and Service Orchestration phases.
• Figure 3: PERFECT's technologies and addressed challenges. The predictive module forecasts UE mobility and request patterns, while the HDRL framework integrates TP, MAC, and PL modules.
• Figure 4: A possible scenario where UE $r_1$ enters from area $a_3$ in time frame $t_1$. Although $u_2$ requires less energy to reach $a_3$ than $u_1$, TP learning algorithm decides to move $u_1$, predicting that $r_2$ will enter from $a_1$. Each grid cell in this deployment across Oulu city represents approximately 2km × 2km.
• Figure 5: Convergence performance of the PERFECT algorithm for different learning rates, highlighting its stability and reward optimization.