A Novel Approach for Spatially Correlated Multi-User MIMO Channel Modeling: Impact of Surface Roughness and Directional Scattering

Recent multiple-input-multiple-output (MIMO) research has moved from independent fading channels to correlated fading channels that better fit real wireless environments. For point-to-point MIMO, many spatially correlated statistical or experimental channel models have been presented in the literature. However, correlated multi-user MIMO channels have not yet been studied, which is critical for investigating the wireless network capacity. Moreover, all previous physical ray tracing approaches assume omni-directional scatterers. However, in real environments, scatterers are not necessarily omni-directional. In this paper, a novel approach for modeling multi-user MIMO channels is proposed. The technique is applicable (in our first attempt) to the non-line-of-sight (NLOS) situation. It considers the impact of scatterers roughness which leads to the directionality in propagation. The directionality of the scatterers is described by the Helmholtz-Kirchihoff field integral and random rough surface theory. In urban street environments, the closed form expression of intra-user (point-to-point MIMO) and inter-user (multi-user MIMO) correlation of any pairwise channel gains is derived. Moreover, the coherence region (where inter-user correlation is expected to be high) is quantitatively presented. It is observed that the coherence region tends to be small for smaller roughness, i.e., higher directionality of the scatterers. This result would help to build realistic fading models for the capacity studies in multiuser MIMO systems.