Reconfigurable intelligent surface (RIS) has emerged as a promising technology for enhancing communication systems. This paper investigates a novel flexible-position RIS-assisted communication system, where the RIS is mounted on a slide rail, enabling spatial adaptability. Unlike traditional fixed-position RISs, the proposed system leverages both spatial flexibility and phase reconfigurability to optimize system performance while reducing overhead through strategic position adjustment. To characterize the spatial variations introduced by RIS movement, we propose a generalized RIS channel model that integrates a practical visibility region function with near-field spherical wave propagation. This model captures the spatial correlation characteristics influenced by multipath angular spread, scatterer distribution, and RIS positioning. Furthermore, we introduce a measurement scheme using a multi-state RIS hardware to analyze segmented channels across fixed-position and flexible-position scenarios. Our measurement reveals that the intra-cluster power angular spectrum follows a Gaussian distribution, and in strong scattering environments, spatial correlation exhibits an enhanced degree of freedom due to spatial non-stationarity effects. In particular, the experimental results demonstrate that the gain of the received power varies from 0.4 dB to 5.3 dB across different RIS positions, providing empirical evidence that spatial adaptability of RIS effectively resists channel non-stationarity. These findings highlight the potential of flexible-position RIS to enhance future wireless communication systems.

Reconfigurable intelligent surfaces (RISs) offer new possibilities for manipulating wireless propagation environments. However, existing studies on RIS-assisted multiple-input multiple-output (MIMO) channel modeling are limited in their ability to explicitly separate and reveal the RIS-induced effects within the cascaded channel. This paper investigates the modeling and measurement of RIS-assisted MIMO channels from a spatial decomposition perspective. We propose a novel stochastic channel modeling framework based on the Weichselberger model, which decomposes the channel into a controllable RIS-induced component and an uncontrollable scattering component. This formulation reveals the RIS’s capability to reshape the spatial coupling structure of MIMO channels and enables the analysis of eigenmode control through spatial correlation and phase configuration. To experimentally validate the model, we develop a channel measurement and separation scheme using a multi-state RIS that can switch among reflective, absorptive, and antenna states. This setup allows for direct extraction of RIS channel component and supports segmented channel measurements. Measurement results show that the RIS-induced and environmental scattering components exhibit distinct power angular spectra and coupling matrices, and their superposition accurately reconstructs the global channel behavior. Additionally, the measured RIS coupling matrix closely matches the one calculated by the theoretical model, with a correlation matrix distance of 0.0394. These results confirm the effectiveness of the proposed model and measurement strategy, offering new insights into RIS-enabled spatial channel customization for next-generation wireless systems.

Reconfigurable intelligent surface (RIS) has gained significant attention as an innovative solution for enhancing communication system performance. This paper proposes a novel fluid RIS-assisted communication system with spatial adaptability, where the RIS is mounted on a slide rail. By leveraging both spatial flexibility and phase reconfigurability, the proposed system overcomes the limitations of conventional fixed-position RIS. To characterize the spatial variations introduced by RIS movement, we establish a novel channel modeling framework that combines near-field spherical wave propagation with practical visibility region functions. Experimental results reveal that dynamic RIS position yields received power gain variations ranging from 0.4 dB to 5.3 dB. This empirical evidence strongly validates the capability of fluid RIS to counteract channel non-stationarity.