Propagation channel characterisation and modelling for high-speed train communication systems

Abstract

High-mobility scenarios, e.g., High-Speed Train (HST) scenarios, are expected to be typical scenarios for the Fifth Generation (5G) communication systems. With the rapid development of HSTs, an increasing volume of wireless communication data is required to be transferred to train passengers. HST users demand high network capacity and reliable communication services regardless of their locations or speeds, which are beyond the capability of current HST communication systems. The features of HST channels are significantly different from those of low-mobility cellular communication systems. For a proper design and evaluation of future HST wireless communication systems, we need accurate channel models that can mimic the underlying channel characteristics, especially the non-stationarity for different HST scenarios. Inspired by the lack of such accurate HST channel models in the literature, this PhD project is devoted to the modelling and simulation of non-stationary Multiple-Input Multiple-Output (MIMO) channels for HST communication systems. In this thesis, we first give a comprehensive review of the measurement campaigns conducted in different HST scenarios and address the recent advances in HST channel models. We also highlight the key challenges of HST channel measurements and models. Then, we study the characterisation of non-stationary channels and propose a theoretical framework for deriving the statistical properties of these channels. HST wireless communication systems encounter different channel conditions due to the difference of surrounding geographical environments or scenarios. HST channel models in the literature have either considered large-scale parameters only and/or neglected the non-stationarity of HST channels and/or only consider one of the HST scenarios. Therefore, we propose a novel generic non-stationary Geometry-Based Stochastic Model (GBSM) for wideband MIMO HST channels in different HST scenarios, i.e., open space, viaduct, and cutting. The corresponding simulation model is then developed with angular parameters calculated by the Modified Method of Equal Area (MMEA). The system functions and statistical properties of the proposed channel models are thoroughly studied. The proposed generic non-stationary HST channel models are verified by measurements in terms of stationary time for the open space scenario and the Autocorrelation Function (ACF), Level Crossing Rate (LCR), and stationary distance for the viaduct and cutting scenarios. Transmission techniques which are capable of utilising Three-Dimensional (3D) spatial dimensions are significant for the development of future communication systems. Consequently, 3D MIMO channel models are critical for the development and evaluation of these techniques. Therefore, we propose a novel 3D generic non-stationary GBSM for wideband MIMO HST channels in the most common HST scenarios. The corresponding simulation model is then developed with angular parameters calculated by the Method of Equal Volume (MEV). The proposed models considers several timevarying channel parameters, such as the angular parameters, the number of taps, the Ricean K-factor, and the actual distance between the Transmitter (Tx) and Receiver (Rx). Based on the proposed generic models, we investigate the impact of the elevation angle on some of the channel statistical properties. The proposed 3D generic models are verified using relevant measurement data. Most standard channel models in the literature, like Universal Mobile Telecommunications System (UMTS), COST 2100, and IMT-2000 failed to introduce any of the HST scenarios. Even for the standard channel models which introduced a HST scenario, like IMT-Advanced (IMT-A) and WINNER II channel models, they offer stationary intervals that are noticeably longer than those in measured HST channels. This has inspired us to propose a non-stationary IMT-A channel model with time-varying parameters including the number of clusters, powers, delays of the clusters, and angular parameters. Based on the proposed non-stationary IMT-A channel model, important statistical properties, i.e., the time-variant spatial Cross-correlation Function (CCF) and time-variant ACF, are derived and analysed. Simulation results demonstrate that the stationary interval of the developed non-stationary IMT-A channel model can match that of relevant HST measurement data. In summary, the proposed theoretical and simulation models are indispensable for the design, testing, and performance evaluation of 5G high-mobility wireless communication systems in general and HST ones in specific.