Beyond Car Sharing: Uncertainty-Aware Pooling of Vehicular Compute at the Network Edge

2026-07-20

AI Summary

This study addresses the challenge of efficiently scheduling computational tasks in highly dynamic and uncertain vehicular environments to leverage vehicle resources as a supplement to edge computing. The authors propose SMART, a novel mechanism within the ETSI MEC framework that, for the first time, integrates calibrated prediction uncertainty into chance-constrained admission control. By employing Bayesian neural networks to forecast future computing capacity and solving the resulting optimization problem via sample average approximation combined with conditional value-at-risk approximation, SMART achieves high performance under uncertainty. Experimental results demonstrate that, using only computing capacity information, the approach attains a task acceptance rate of 95.6% and a median capacity violation rate of merely 0.81%, significantly outperforming existing baselines in the trade-off between admission and violations, and closely approaching the performance of an ideal capacity oracle.

Abstract

Connected vehicles increasingly embed AI accelerators, offering a substantial yet volatile source of supplemental compute near the network edge. Unlike provisioned MEC hosts, vehicular resources are highly dynamic: vehicles may leave the cell, become locally occupied, or offer heterogeneous compute capacities. Therefore, exploiting vehicular resources requires making admission decisions without knowing the compute capacity that will be available during task execution. We present SMART, an uncertainty-aware admission mechanism that enables an \acs{ETSI} \ac{MEC} orchestrator \textit{to opportunistically exploit vehicular compute under predictive uncertainty.} SMART predicts future vehicular capacity using a \ac{BNN} -- whose uncertainty estimates are the best calibrated among the evaluated forecasters at the nominal 95\% level, and incorporates its calibrated predictive uncertainty into a chance-constrained admission program reformulated through \ac{SAA} and \ac{CVaR} approximations. Under a compute-only admission model, SMART admits 95.6\% of tasks while maintaining a median capacity-violation rate of about 0.81\%. It achieves a favorable admission-violation tradeoff compared with seven reactive, mean-only, and uncertainty-aware baselines, by approaching the performance of a compute-capacity oracle under the modeled assumptions. Finally, the sensitivity analyses show that variability in base-station compute availability is a key determinant of admission performance, highlighting the need for a calibrated admission and resource allocation mechanism tailored to opportunistic base-station compute pooling.

Problem

• vehicular compute
• predictive uncertainty
• edge computing
• resource pooling
• admission control

Innovation

• uncertainty-aware admission
• Bayesian neural network
• vehicular edge computing
• chance-constrained optimization
• opportunistic compute pooling