Delay-Phase Joint Optimization for Wideband Hybrid Precoding Massive MIMO Systems with Angular Spread
摘要
Terahertz (THz) communications are considered a core technology for beyond-5G and 6G networks, offering ultra-wide bandwidths but facing significant challenges due to severe path loss, beam squint, and angular spread sensitivity. This paper proposes a novel wideband hybrid precoding framework for multi-user THz massive MIMO (Multiple-Input Multiple-Output) systems, where the base station employs a Uniform Planar Array (UPA) to support 3D beamforming. The system model incorporates a THz-specific wideband channel, accounting for frequency-selective fading and molecular absorption. The proposed approach formulates the hybrid precoding problem as a temporal-phase matrix decomposition, supporting multiple implementations including DP-AltMin, TTD (True Time Delay) assisted DPP, and MO-AltMin architectures. Extensive simulations demonstrate that the proposed DP-AltMin precoding achieves a 20.6% improvement in sum spectral efficiency over spatially sparse precoding and a 13.8% gain over TTD-assisted hybrid schemes under an angular spread of