부산대학교 도서관

Precise modelling of the Direction of Arrival (DoA) and Direction of Departure (DoD) of multipath components (MPCs) in a Multiple-Input Multiple-Output (MIMO) channel based on correlation properties and joint distributions have not been investigated properly. This paper examines the correlation between DoAs and DoDs, and presents a modelling approach particularly for the `general trend' of these dual-spatial domains, based on statistical results obtained from several indoor MIMO measurement campaigns. We show that the Laplacian distribution provides better fits for the directional parameters than the normal and Von Mises distributions. The `general trend' of DoA/DoD joint distribution and power density are modelled using a modified Gumbel's bivariate exponential approach. Results show that the proposed bivariate distribution provides a good match to the real data, and offers an alternative approach to generate realistic directional-based stochastic MIMO channel responses in different environments.