Proximal Policy Optimization for RIS-assisted Full Duplex 6G-V2X Communications

Abstract

In this work, we consider a novel reconfigurable intelligent surface (RIS)-assisted full-duplex (FD) sixth generation (6G)-vehicle-to-everything (V2X) communication network having a FD base station (BS) simultaneously communicating with a uplink (UL) and a downlink (DL) mobile vehicles with the aide of two RISs, one for each link. We provide an analytical framework to investigate the performance of this network and, consequently, formulate an optimization problem to jointly optimize the phase shift matrices at both the RISs that maximize the achievable sum-rate. Thereafter, due to the non-convex nature of the problem, we propose a low complexity proximal policy optimization (PPO)-based deep reinforcement learning (DRL) algorithm that solves the problem in continuous action spaces by reducing the overall training overhead and provides optimum values of phase-shift matrix at each RIS. Further, we extend the analysis considering multiple transmit and receive antennas at the BS that simultaneously serves multiple UL and DL vehicles. We present exhaustive simulations-based graphical results to validate the effectiveness and accuracy of the proposed algorithm compared to deep deterministic policy gradient (DDPG) and successive refinement (SR)-based solutions. Accordingly, we demonstrate the dominance of the considered FD system over its half-duplex (HD) counterpart. We also discuss the impact of the number of elements at each RIS, each vehicle's velocity and their distance from the RIS and the BS on the performance of the respective links. Moreover, we also highlight the impact of imperfect self-interference cancellation and discuss the trade-off between the UL and DL performances due to this imperfection. Additionally, it is shown that the proposed algorithm is approximately <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$10 \times$</tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$ 40 \times$</tex-math></inline-formula> faster than DDPG and SR, respectively. Finally, for the multiple vehicle case, we demonstrate the effect of co-channel interference on the DL rate and the impact of RIS elements, the number of UL and DL vehicles on the system performance.