RIS-Assisted MIMO Channel Modeling with Spatially Consistent Sparsified Properties

Semyon Dorokhin,Pavel Lysov,Vladimir Lyashev,Andrei Kunavin,Anastasia Aderkina
DOI: https://doi.org/10.1007/s11277-024-10884-z
IF: 2.017
2024-03-27
Wireless Personal Communications
Abstract:Reconfigurable Intelligent Surface is a planar array that implements partially controllable propagation environment concept. RIS quickly gained popularity in research community for its potential in 6 G applications. However, most of the researchers who analyze RIS-assisted systems use simple Rician channel model. Being a convenient tool for theoretical analysis, Rician model does not provide realistic Non Line-of-Sight (NLoS) modeling. Furthermore, Rician model is not eligible to evaluate 3GPP standardization proposals. That is why in this article we propose a Geometry-Based Stochastic Model for RIS-assisted systems. Our model is based on 3GPP-compliant channel models from the popular QuaDRiGa simulation platform. We compare the proposed model against Rician and simplified 3GPP model from literature in MIMO scenario with perfect CSI assumption and in MISO scenario with both perfect CSI and Compressed-Sensing based channel estimation. The comparison reveals that the difference in direct channel NLoS modeling is the main contributor to the difference between the chosen models: the capacity gain with Rician channel is twice smaller compared to 3GPP-based models. If the RIS is large and the cascaded channel is much stronger than the direct, the average achievable rate of MISO RIS-aided system differs only by . For small RIS however, there can be up to difference in achievable rate. We also demonstrate that the proposed model has almost 10 times larger achievable rate variance, which is critical if Outage Probability is considered.
telecommunications
What problem does this paper attempt to address?