Statistical millimeter wave channel modelling for 5G and beyond

Abstract

Millimetre wave (mmWave) wireless communication is one of the most promising technologies for the fifth generation (5G) wireless communication networks and beyond. The very broad bandwidth and directional propagation are the two features of mmWave channels. In order to develop the channel models properly reflecting the characteristics of mmWave channels, the in-depth studies of mmWave channels addressing those two features are required. In this thesis, three mmWave channel models and one beam alignment scheme are proposed related to those two features. First, for studying the very broad bandwidth feature of mmWave channels, we introduce an averaged power delay profile (APDP) method to estimate the frequency stationarity regions (FSRs) of channels. The frequency non-stationary (FnS) properties of channels are found in the data analysis. A FnS model is proposed to model the FnS channels in both the sub-6 GHz and mmWave frequency bands and cluster evolution in the frequency domain is utilised in the implementation of FnS model. Second, for studying the directional propagation feature of mmWave channels, we develop an angular APDP (A-APDP) method to study the planar angular stationarity regions (ASRs) of directional channels (DCs). Three typical directional channel impulse responses (D-CIRs) are found in the data analysis and light-of-sight (LOS), non-LOS (NLOS), and outage classes are used to classify those DCs. A modified Saleh-Valenzuela (SV) model is proposed to model the DCs. The angular domain cluster evolution is utilised to ensure the consistency of DCs. Third, we further extend the A-APDP method to study the spherical-ASRs of DCs. We model the directional mmWave channels by three-state Markov chain that consists of LOS, NLOS, and outage states and we use stationary model, non-stationary model, and “null” to describe the channels in each Markov state according to the estimated ASRs. Then, we propose to use joint channel models to simulate the instantaneous directional mmWave channels based on the limiting distribution of Markov chain. Finally, the directional propagated mmWave channels when the Tx and Rx in motion is addressed. A double Gaussian beams (DGBs) scheme for mobile-to-mobile (M2M) mmWave communications is proposed. The connection ratios of directional mmWave channels in each Markov state are studied.