Max-Min Secrecy Rate and Secrecy Energy Efficiency Optimization for RIS-Aided VLC Systems: RSMA Versus NOMA

TL;DR

The paper addresses secure downlink transmission in RIS-aided VLC systems by jointly optimizing VLC AP power, RIS association, and RIS element orientations under RSMA and NOMA. It formulates max-min secrecy rate and secrecy-energy-efficiency problems and solves them with a genetic algorithm, handling mixed-integer variables and random device orientation. Results show RSMA consistently outperforms power-domain NOMA across SR and SEE metrics, with insights into how RIS element count, reflectivity, and minimum-rate requirements influence performance. The work demonstrates the practical viability of GA-based joint optimization for secure indoor optical wireless networks and suggests avenues for scaling and robustness improvements in future research.

Abstract

Integrating VLC with the RIS significantly enhances physical layer security by enabling precise directional signal control and dynamic adaptation to the communication environment. These capabilities strengthen the confidentiality and security of VLC systems. This paper presents a comprehensive study on the joint optimization of VLC AP power allocation, RIS association, and RIS elements orientation angles for secure VLC systems, while considering RSMA and power-domain NOMA schemes. Specifically, two frameworks are proposed to maximize both the minimum secrecy rate (SR) and the minimum secrecy energy efficiency (SEE) by jointly optimizing power allocation, RIS association, and RIS elements orientation angles for both power-domain NOMA and RSMA-based VLC systems. The proposed frameworks consider random device orientation and guarantee the minimum user-rate requirement. The proposed optimization frameworks belong to the class of mixed integer nonlinear programming, which has no known feasible solution methodology to guarantee the optimal solution. Moreover, the increased degree of freedom and flexibility from the joint consideration of power control, RIS association and element orientation results in a large set of decision variables and constraints, which further complicates the optimization problem. To that end, we utilize a genetic algorithm-based solution method, which through its exploration and exploitation capabilities can obtain a good quality solution. Additionally, comprehensive simulations show that the RSMA scheme outperforms the power-domain NOMA scheme across both the SR and SEE metrics over various network parameters. Furthermore, useful insights on the impact of minimum user rate requirement, number of RIS elements, and maximum VLC AP transmit power on the minimum SR and SEE performances are provided.

Figures (9)

• Figure 1: (a) An illustration of the proposed mirror array-based -aided secure system, (b) the angles of irradiance and incidence of the light signal that propagates from the through the $k$-th mirror array-based element to the $u$-th user, (c) the eavesdropped light signal while specifying the angles of incidence and irradiance of this signal, (d) the user device's orientation in respect of the device's polar and azimuth angles ($\alpha_u,\beta_u$), (e) the mirror array-based orientation in respect of its roll angle $\omega_k$, and (f) the mirror array-based orientation in respect of its yaw angle $\gamma_k$.
• Figure 2: Max-Min performance versus electrical transmit power for different .
• Figure 3: Max-Min performance versus number of elements for different reflectivities.
• Figure 4: Max-Min performance versus users' minimum rate requirement.
• Figure 5: Max-Min performance versus number of intended users.