Reduced-complexity cluster modeling for time-variant wideband MIMO channels

This paper presents a reduced-complexity cluster modeling method for channel models that are based on the 3GPP and similar channel models to simulate the time variation of spatially correlated wideband MIMO channels. The main novelty is that, when modeling the time-variant wideband MIMO channels, instead of tracking the changes in the angles of arrival (AoAs) of all the multi-path components (MPCs) defined, we only track the change in the centre AoA for each of the clusters. Hence for moderate angle spreads (ASs) of clusters and a constant uniform distribution of the offsets of the MPCs within each cluster, tracking the time-variant centre AoAs of the clusters allows us to develop a computationally efficient approximation method to calculate the instantaneous channel matrix and spatial correlation matrix for time-variant wideband MIMO channels. The development of this approximation method includes two stages: firstly, we evaluate the approximation method for simulating wideband MIMO channels with time-invariant AoAs in terms of the centre angles and scatterer distributions of clusters; secondly, on the basis of the validation at stage one, we develop the approximation method for the wideband MIMO channels with time-variant AoAs, and evaluate this approximation method by the extended correlation matrix distance (CMD) metric. We use the extended CMD metric to compare the CMDs predicted by the approximate and exact calculation under different time-variant scenarios. The simulation results show that the approximation method works well when the velocity of the movement is up to 50 m/s and provided the ASs of the clusters are within 10∘.