Abstracting indoor signal propagations: Stochastic vs. regular

In this paper, we investigate the performance of indoor users by introducing two abstraction models for the indoor environment, which influence the signal propagation. On the one hand, we discuss a mathematically tractable approach that is based on randomly distributed walls according to a random object process (ROP). The opposing approach is realized by generating a Manhattan grid with equidistantly spaced vertical and horizontal walls. In order to provide a fair comparison, we assure the same wall volume in both scenarios. This volume is calculated based on the wall density and the average wall length in the ROP. We investigate the performance of a cell-edge user in a setup comprising four transmitters, applying numerical evaluations and system level simulations. The results for the average attenuation due to blockage exactly overlap for both wall generation approaches, and also show a perfect match with the analysis. In contrast, the signal-to-interference ratio and the corresponding throughput results obtained in an LTE-A system diverge.