In order to fully exploit the potential that the multiple-input multiple-output (MIMO) technology can provide for the novel radio communication applications, knowledge of the radio channel is necessary. For instance, signal processing algorithms or network coverage planning are tasks that are vitally dependent on the characteristics of the radio channel in which the system is desired to operate. However, since it is both time-consuming and expensive to measure all the envisioned usage scenarios, accurate and easy-to-use channel models are essential in many stages of the system development. This thesis aims at improving the quality of the geometry-based stochastic MIMO channel models (GSCMs). First, an overview of the existing MIMO channel models is given including a detailed description of the principles of the models using the geometry-based approach. In addition, the shortages of the current GSCMs are discussed in order to motivate the work of the thesis on their part. The main achievements of this thesis are the following. First of all, as compulsory background work, a measurement-based ray tracer (MBRT) was developed in order to facilitate detailed analysis of the radio channel measurements. With the help of the MBRT, channel model parameters for GSCMs were extracted from measurement data gathered in various indoor environments. In addition, the characteristics of the so called dense multipath components (DMC) were comprehensively studied, and as a result, a method to include the DMC to the GSCMs was developed. Finally, issues related to multi-link MIMO channel modeling were addressed. First and foremost, the propagation phenomena that are important in multi-link scenarios were studied. Based on the analyses, an approach to extend current GSCMs to fully support simulations of multi-link scenarios was invented. Many of the outcomes of this thesis have been directly applied in the COST 2100 MIMO channel model.
展开▼
