SLNR-Based Private Precoding for RSMA Systems: Analysis and Performance Evaluation

Abstract

This paper investigates signal-to-leakage-and-noise ratio (SLNR) precoding for private streams in the downlink of a multiple-input single-output (MISO) rate-splitting multiple access (RSMA) system. The common stream is designed using a multistart gradient-ascent method, whereas the private precoders are obtained analytically using SLNR. For comparison, the analytical zero-forcing (ZF) method and the numerical iterative successive null-space (SNS) and weighted minimum mean-square error (WMMSE) methods are also considered for private-stream precoding. The study accounts for imperfect CSIT, spatial correlation, and different numbers of base-station antennas. Simulations are conducted under an equal-rate strategy for different SNR and correlation levels. The results indicate that SLNR outperforms ZF and SNS in the correlated channels considered and is more robust to CSI errors. Although SLNR performs slightly below WMMSE, it has substantially lower computational complexity. The results demonstrate that SLNR is a practical, low-complexity solution for private-stream precoding in next-generation RSMA systems, including 6G scenarios.