Deep Learning-Based Tri-Hybrid Multi-User MIMO Precoding: The Blessing of EM-Reconfigurable Antennas
TriPNet is a tri-hybrid precoding network that learns electromagnetic (EM), analog, and digital precoding simultaneously from channel data to boost spectral efficiency in wideband multi-user MIMO systems. It uses a Conformer encoder for pattern selection and combines cross-attention with SVD/ZF priors for hybrid precoding. This approach outperforms existing methods while maintaining low online complexity.