Abstract
The large number of hardware-efficient elements incorporated in reconfigurable intelligent surface (RIS) brings benefits in low power consumption and high throughput in wireless communications, however, such merits come along with challenges in channel estimation (CE). Classic CE approaches are ineffective and hardware-costly, thus, an effective CE method with low overhead is required. As an evolution of RIS, holographic multiple-input single-output surface (HMIMOS) is able to form a continuous aperture, boosting great potential in communications. However, such novel reconfigurable technology is also accompanied by technical issues, i.e., channel modeling and performance analysis in far-field and near-field regions. To deal with these challenges, this thesis investigated RIS-assisted wireless communications and HMIMOS systems separately. Primarily, for RIS-assisted communications, the tensor decomposition-based architecture is exploited for CE, which adopts the parallel factor decomposition to unfold the resulting cascaded channel model with alternating least squares and vector approximate message passing algorithms. To further enhance CE performance and increase efficiency in RIS-empowered wireless communications, two joint CE and signal recovery schemes based on message passing algorithms are examined in this thesis. The simulation results prove that these CE schemes can reduce complexity and achieve favorable both CE performance and detection performance without compromising the robustness and convergence of algorithms. Additionally, regarding HMIMOS systems, this thesis proposesEMcompliant channel modeling for far-field and near-field communications in two separate sections. The far-field channel model is based on Fourier expansion in wavenumber domain while the near-field channel model is based on dyadic Green’s functions. The hardwareefficient precoding schemes are presented to eliminate spatial (or polarization) interferences. The theoretical correlation analysis is also derived to support EM interpretations. The simulation results all substantiate the validity of the proposed channel models, providing insights into system design and practical applications.