Reconfigurable intelligent surface (RIS) is considered as a promising technology for future wireless communication, localization and sensing systems. An RIS is a thin planar array that consists of multiple sub-wavelength-sized reflecting elements, each of which can induce independent and controllable interactions with the incident signal. An RIS can actively control the propagation channel, creating energy-efficient smart radio environments. Meanwhile, polarization is a fundamental property of electromagnetic (EM) waves; it refers to the orientation of the electric field component relative to the direction of propagation. The EM polarization provides additional degrees of freedom (DoF) to the wireless channel beyond the well-exploited time, frequency, and spatial dimensions.

In this thesis, we explore the RIS technology and EM polarization as two potential key enablers for future wireless communication and localization systems. First, we exploit the RIS technology to address a fundamental bottleneck for point-to-point multiple-input multiple-output (MIMO) communications in line-of-sight environments. Specifically, we deploy multiple RISs as controllable scatterers to create an artificial scattering environment, thereby providing additional spatial DoF and restoring the spatial multiplexing gain of MIMO systems.

Second, we present two studies that focus on the RIS operation regarding its efficient position of deployments and the channel state information acquisition. In the former study, we analyze the positional impact of the RIS on the achievable rate for single and multiple antenna systems. We consider not only the impact of propagation losses but also the impact of channel richness with multipath components to identify the efficient region of RIS deployments. In the latter study, we develop two distinct channel estimation schemes based on the superimposed pilot signals to reduce the significant pilot overhead required in RIS-aided wideband multi-user communications. Then, we incorporate iterative channel estimation and data detection techniques. Furthermore, we develop an RIS reflection design using an alternating optimization to maximize channel power.

Third, we unlock the RIS capability to control the polarization states of reflected signals by considering dual-polarized RIS. We utilize the RIS as an information modulator by performing three distinct polarization-based modulation schemes: binary polarization modulation, joint differential polarization and phase modulation, and joint polarization and spatial modulation. 

Finally, we study the potential benefits of integrating the polarization dimension into localization applications. We identify the DoF provided by polarization for localization. Subsequently, we leverage these DoF to introduce three distinct localization applications which enable three dimensional (3D) orientation estimation, 2D angle of departure estimation, and mixed 2D position and 1D orientation estimation for vehicular scenarios.