Communication-Aware Consistent Edge Selection for Mobile Users and Autonomous Vehicles

TL;DR

This paper tackles the challenge of maintaining low-latency, time-sensitive offloading for mobile vehicles by addressing frequent AP handovers and edge migrations in vehicular MEC. It introduces a Deep Deterministic Policy Gradient (DDPG) based framework that jointly optimizes AP-edge selection and computation offloading, incorporating SNR and AP-load into the decision process to balance link quality with stability. An MDP formulation with a carefully designed state, action, and reward structure guides offline training to learn robust AP-vehicle associations, prioritizing high QoS while minimizing handovers. Simulation results in Gazebo/ROS2 demonstrate that the proposed method reduces handovers and latency while maintaining high SNR and balanced edge-load, outperforming RA, SSF, and LLF baselines and highlighting its practical potential for dynamic vehicular edge computing in 5G/V2X environments.

Abstract

Offloading time-sensitive, computationally intensive tasks-such as advanced learning algorithms for autonomous driving-from vehicles to nearby edge servers, vehicle-to-infrastructure (V2I) systems, or other collaborating vehicles via vehicle-to-vehicle (V2V) communication enhances service efficiency. However, whence traversing the path to the destination, the vehicle's mobility necessitates frequent handovers among the access points (APs) to maintain continuous and uninterrupted wireless connections to maintain the network's Quality of Service (QoS). These frequent handovers subsequently lead to task migrations among the edge servers associated with the respective APs. This paper addresses the joint problem of task migration and access-point handover by proposing a deep reinforcement learning framework based on the Deep Deterministic Policy Gradient (DDPG) algorithm. A joint allocation method of communication and computation of APs is proposed to minimize computational load, service latency, and interruptions with the overarching goal of maximizing QoS. We implement and evaluate our proposed framework on simulated experiments to achieve smooth and seamless task switching among edge servers, ultimately reducing latency.

Figures (10)

• Figure 1: Communication-aware edge selection in the multi-edge system for a fleet of autonomous vehicles, aided by a DDPG-based learning framework.
• Figure 2: DDPG-based Communication-aware Edge Selection Framework for a Fleet of Autonomous Vehicles.
• Figure 3: The simulation setup shows Scenario 1 (left) and Scenario 2 (right) with different access point locations, as well as the starting and ending (goal) points of the autonomous e-vehicles in the simulation trials.
• Figure 4: Convergence of the proposed approach based on training performed on simulations and the parameters of Table II.
• Figure 5: Performance of the critic loss function during training performed on simulations and the parameters of Table II.