Joint vehicular power and RIS reflection coefficient optimization to maximize the minimum bit rate in VANETs

Abstract

Reconfigurable intelligent surfaces (RISs) have recently gained significant attention for improving reliability in vehicular communications. However, ensuring reliable communication between far-distance vehicles remains a challenge. This work investigates an RIS-aided vehicular ad hoc network (VANET) in a road section where the distance between vehicles is too large for a single roadside unit (RSU) to provide reliable communication. We propose a novel RIS architecture where several RIS panels are connected by cables and each is equipped with a power amplifier. We then optimize vehicle power and RIS reflection coefficients to maximize the minimum bit rate of the VANET. Due to the non-convex nature of the formulated problem, we use fraction programming (FP) to reformulate it into a convex form, allowing solution using tools like CVX which is a MATLAB-based modeling system for convex optimization. The reformulated problem is then decoupled into subproblems. Block coordinate descent (BCD) is employed to optimize all variables alternately and obtain the joint optimal solution. Additionally, the alternating direction method of multipliers (ADMM) ensures that the phase shift of each reflecting element remains a unit vector. Finally, semidefinite relaxation (SDR) is used to solve the boolean quadratically constrained problem. Simulation results demonstrate the effectiveness of the proposed method and confirm that our architecture outperforms conventional approaches.