ABSTRACT This paper develops novel precoding techniques tailored for massive multiuser (MU) MIMO channels, utilizing a supervised singular value decomposition (SupSVD) approach at the precoding stage. These techniques effectively represent the channel's randomly varying data, significantly minimizing noise interference. By accurately depicting channel behaviors in ultra‐dense network environments, the proposed methods enhance system reliability. High‐density networks often suffer from severe interference and limited spectral efficiency; our MU‐MIMO solutions address these challenges by optimizing signal clarity and reception quality across numerous simultaneous transmissions. We apply our precoding solutions to Rayleigh fading MU‐MIMO channels across downlink and uplink configurations. Comparative simulation studies reveal that our methods outperform traditional linear precoding techniques, demonstrating a notable 3 dB improvement in bit error rate (BER). However, the complexity of the precoding stage can introduce computational challenges, requiring careful algorithm optimization to balance performance with practical feasibility.
Abdelbari et al. (Tue,) studied this question.