Spatial Consistency of 3D Channel Models

Developing realistic channel models is one of the greatest challenges for describing wireless communications. Their quality is crucial for accurately predicting the performance of a wireless system. While on the one hand, channel models have to be accurate in describing the physical properties of wave propagation, on the other hand, they have to be as least complex as possible. With the recent emergence of antennas with a massive amount of elements as a promising technology for a further enhancement of spectral efficiency, new channel models that characterize the propagation environment in both azimuth and elevation become necessary. While standardization bodies such as 3rd Generation Partnership Project (3GPP) and International Telecommunications Unit (ITU) have introduced a 3-dimensional (3D) geometry-based stochastic channel model, a system-level modeling has been missing to serve the purpose of further analysis and evaluations. Furthermore, with such a channel characterization, where both geometry and statistics are jointly combined, it is crucial that spatial consistency is included in the model. Facing the challenge of channel parameters being both position- and time-dependent, as well as the lack of spatial consistency, this dissertation presents a system-level framework and design of the 3D geometry-based stochastic channel models enhanced with spatial consistency. In the first part of this dissertation, the design of the 3D geometry-based stochastic channel models (GSCMs) on an existing system-level tool is considered. The focus is put on modeling aspects of large-scale and small-scale fading being both position- and time-dependent, a challenge for the already complex structure of system-level tools. A novel design is proposed for spatial granularity and time line structure of simulation tools that reduces the simulation complexity and enables the generation of the channel impulse response at runtime. Furthermore, the proposed design facilitates the key functionality of wireless communications, mobility of the user and a time evolution of the channel impulse response while the user is on the move. It is shown that this structure is a key element in shaping the design of the upcoming fifth generation (5G) and beyond system-level tools. Modeling of the spatial consistency is discussed in the second part of this dissertation. Two novel models are proposed: The first one establishes spatial correlation properties among propagation conditions such as line of sight (LOS non line of sight (NLOS), indoors and outdoors based on 2- dimensional spatial filtering. Further it is shown that this model yields a realistic behavior mimicking the effects of blockages in 3D. The second model introduces spatially consistent small-scale fading. A correlation among random variables is proposed based on a specific resolution, namely the decorrelation distance, that indicates the range in which random variables are independent. Further, the model is validated by comparing to statistical measures extracted from extensive ray-tracing simulations. The proposed model is in a very good agreement with statistical channel properties obtained from ray-tracing. At the end, the proposed model is parametrized, and based on hypothesis testing the de-correlation distance values for various scenarios are determined.