Reconfigurable Intelligent Computational Surfaces for MEC-Assisted Autonomous Driving Networks: Design Optimization and Analysis
Xueyao Zhang, Bo Yang, Zhiwen Yu, Xuelin Cao, George C. Alexandropoulos, Yan Zhang, Merouane Debbah, Chau Yuen
TL;DR
The paper tackles safety in autonomous driving by offloading sensing tasks from CVs to MEC via V2I while mitigating interference on V2V. It introduces Reconfigurable Intelligent Computational Surfaces (RICS) with RR+AC mode to enable reflective V2I communication and analog computing for interference suppression. A joint optimization framework (AIOA) optimizes task offloading ratios ${\rho_m}$, spectrum sharing ${\alpha_{m,n}}$, and RICS reflection/refraction matrices, solved by decomposing into three subproblems and applying FP_Quad, SCA, SDR, and gradient-descent methods. Numerical results show significant safety gains (up to ~34% in safety coefficient) and around a 60% improvement in V2V data rate, demonstrating the practical impact of RICS-enabled MEC in dense vehicular networks.
Abstract
This paper investigates autonomous driving safety improvement via task offloading from cellular vehicles (CVs) to a multi-access edge computing (MEC) server using vehicle-to-infrastructure (V2I) links. Considering that the latter links can be reused by vehicle-to-vehicle (V2V) communications to improve spectrum utilization, the receiver of the V2I link may suffer from severe interference that can cause outages during the task offloading. To tackle this issue, we propose the deployment of a reconfigurable intelligent computational surface (RICS) whose computationally capable metamaterials are leveraged to jointly enable V2I reflective links as well as to implement interference cancellation at the V2V links. We devise a joint optimization formulation for the task offloading ratio between the CVs and the MEC server, the spectrum sharing strategy between V2V and V2I communications, as well as the RICS reflection and refraction matrices to maximize an autonomous driving safety task. Due to the non-convexity of the problem and the coupling among its free variables, we transform it into a more tractable equivalent form, which is then decomposed into three sub-problems solved via an alternate approximation method. Our simulation results showcase that the proposed RICS-assisted offloading framework significantly improves the safety of the considered autonomous driving network, yielding a nearly 34\% improvement in the safety coefficient of the CVs. In addition, it is demonstrated that the V2V data rate can be improved by around 60\% indicating that the RICS-induced adjustment of the signals can effectively mitigate interference at the V2V link.