Joint Precoding and Beamforming for Active Beyond-Diagonal RIS-Assisted Multi-User DFRC
TL;DR
The proposed architecture achieves superior radar beampattern control and favorable energy efficiency-performance trade-offs, establishing active BD-RIS as a promising enabler for sensing-aware, energy-efficient, and high-capacity 6G DFRC networks.
Abstract
This paper investigates the integration of active beyond-diagonal reconfigurable intelligent surfaces (BD-RIS) into dual-functional radar-communication (DFRC) systems for multi-user sixth generation (6G) networks. Unlike conventional passive or diagonal RIS, the proposed architecture employs active non-reciprocal impedance networks that enable joint amplitude and phase control, thereby mitigating multiplicative fading and enabling direction-dependent wave manipulation. We formulate a joint optimization problem to maximize the weighted sum rate (WSR) while preserving radar probing capability, subject to base-station transmit power and RIS amplification constraints. To solve the resulting non-convex problem, we develop an alternating optimization framework combining weighted minimum mean-squared error (WMMSE)-based precoder design, fractional programming for active BD-RIS beamforming, and quasi-Newton projection to enforce group-wise amplification constraints. The computational complexity of the proposed algorithm is rigorously characterized, highlighting the scalability benefits of group-connected BD-RIS architectures. Extensive simulations demonstrate up to 2.8 bps/Hz WSR improvement over passive diagonal RIS, 41.75% transmit power reduction compared with active diagonal RIS, and strong robustness under imperfect channel state information and hardware impairments. Furthermore, the proposed architecture achieves superior radar beampattern control and favorable energy efficiency-performance trade-offs, establishing active BD-RIS as a promising enabler for sensing-aware, energy-efficient, and high-capacity 6G DFRC networks.