Beyond-diagonal RIS and hybrid NOMA/OMA: A fairness-driven framework for collaborative industrial devices

Emerging wireless technologies are redefining dense Industry 5.0 ecosystems, where fairness, low latency, and adaptive connectivity are critical for supporting heterogeneous, resource-constrained Collaborative Industrial Devices (CIDs) such as autonomous guided vehicles (AGVs), robotic actuators, and industrial sensors. This paper presents a fairness-aware resource management framework for industrial networks that combines Beyond-Diagonal Reconfigurable Intelligent Surfaces (BD-RIS), centralized orthogonal multiple access (OMA), and non-orthogonal multiple access (NOMA)-enabled edge connectivity. A joint max-min rate optimization problem is formulated to manage device association, subcarrier allocation, access mode selection, power control, decoding order, and RIS phase configuration. To solve the resulting mixed-integer non-convex problem, we develop a block coordinate descent (BCD) algorithm leveraging convex approximation, Riemannian optimization, and successive interference cancellation (SIC)-aware clustering. Simulations show that the proposed scheme achieves a minimum rate of 1.88 bps/Hz at 32 dBm, improving by 42% and 39% over RIS-OMA and full NOMA with random RIS phases (FNOMA-RR) baselines, respectively. The framework scales well to high CID densities, maintains fairness (median rate: 1.45 bps/Hz, narrow interquartile range), and achieves robust load balancing across access points. These results confirm the practical feasibility of our approach in enabling reliable, fair, and spectrally efficient downlink connectivity for next-generation Industry 5.0 environments.