Research On Wireless Channel Measurement And Modeling In Complex Scenarios Of Vehicular Networks

Vehicular wireless communications use land moving vehicles as communication ter-minal, transmit signals via wireless network, and dynamically exchange the safety and service information of users. The generalized vehicular communications include rail and road traffics. With the development of the Intelligent Transportation System (ITS), the future hybrid transportation, which combines both rail and road traffics, will be an impor-tant approach to improve transportation efficiency and ensure safety.The development of ITS relies on vehicular wireless communications, and the design of wireless network requires a deep investigation of wireless channels in typical vehicular environments. As the physical wireless channel is the medium over which communica-tion happens, its properties significantly affect the performance of wireless communica-tion system. In the design of vehicular wireless communication system, channel models are required to describe the complex vehicular propagation channels. Even though nu-merous research has been conducted on the vehicular channels, however, there are still the following limitations:1) In terms of scenario partitioning, the current research only covers a few parts of vehicular propagation scenarios due to the limitation of measurements. It fails to meet people’s demand for understanding propagation in vehicular scenarios, and can not provide required radio channel models the system designers.2) In terms of scientific research, the current research on vehicular wireless channels fails to give insights into the fundamental law of radio wave propagation. There lacks deep investigations on definition of channel non-stationarity in moving sce-narios, fading statistics of time-variant channels, and theory of dynamic channel modeling. The current limitations on scientific level could lead to potential risk to reliability of system.3) In terms of applications, the current vehicular channel models can not meet the demands for link budget, simulation, and performance validation in system design. The current models can neither cover typical vehicular scenarios, nor reflect time-varying characteristics of vehicular channels. These limitations significantly reduce accuracy of system simulation. In order to solve above problems, this dissertation investigates the radio propagation in high-speed railway and vehicle-to-vehicle environments, and proposes channel models for vehicular network design. The main works are as follows:1) For the complexity of high-speed railway environments, a scene partitioning stan-dard is proposed. Based on propagation breakpoint estimation under directional base station antennas and pattern calibration, a large-scale propagation model base is developed for high-speed railways, including path loss prediction model, shadow fading distribution and correlation models. The proposed model provides scientific references for link budget of system.2) For the time-variant fading channels of high-speed railways, the small-scale fad-ing characteristics in viaduct and cutting scenarios are investigated. The impact of scenario structures on distribution of multi-path components is discussed. A dis-tance and scenario structure based piecewise linear model of distribution parameter is proposed for the time-variant channels. The model gives insight into fast fading channels in high-speed railway environments.3) For the complex Non-Line-of-Sight (NLOS) scenarios in vehicle-to-vehicle chan-nels, two typical scenarios, street large vehicular obstructions and cross-road, are investigated. A double distance decay factors based path loss model is proposed for the scenarios with large vehicular obstructions, and the screen effect of cross-road bend and roadside obstacles in cross-road scenario is studied based on distribution of multi-path components on angular domain.4) A metric of estimating wide-sense stationarity region for time-variant channels is proposed based on the correlation matrix distance. The fundament of non-stationary channels, which is the random "birth" and "death" processes of multi-path compo-nents, is fully discussed based on angular-time-frequency joint spectrum estimation. A multi-path component identification and tracking algorithm is proposed, and a dynamic wideband directional channel model is developed for vehicle-to-vehicle communications, which helps to enrich the theory of dynamic channel modeling.To sum up, this thesis investigates the wireless channels in high-speed railway and vehicle-to-vehicle environments, supports propagation prediction tools for the system de-signs of high-speed railway and vehicle-to-vehicle communications, compensates the d- eficiency of channel modeling in vehicular complex scenarios, and enriches the theory of modeling in such scenarios.