Modeling And Simulation Of High Mobility Wireless Channels Based On Xiao Model And Basis Expansion Models

Under the highly mobile broadband communication scenarios, the wireless propagation environment becomes very complex. The channel will experience the time-frequency doubly-selective fading. In order to study the modulation, channel estimation and equalization techniques for high-speed railway broadband wireless communication systems, the wireless channel modelling becomes very important.The thesis firstly discusses the characteristics of the wireless channel model of high-speed railway, including the wireless channel path loss, Doppler effect, multipath effect, and the corresponding fast fading, slow fading, flat fading, frequency selective fading and high-speed railway penetration loss. On this basis, the thesis analyzes the flat fading and frequency selective fading modeling and simulation methods. The simulation of flat fading uses improved Jakes model (Xiao model) and the simulation of frequency selective fading uses TDL (Tapped Delay Lines) structure. The thesis also briefly discusses the principles and system architecture of the OFDM system.Then, based on the idea of Qian Liu et al, the thesis proposes a fading simulation model suitable for high-mobility Rice fading and analyzes the autocorrelation function of the phase component, cross-correlation function, square envelope autocorrelation function and probability density functions under the constant speed and accelerated operation. The simulation results show that the improved model can not only be used for the train moving with a constant speed, but also be used for the train in the acceleration process. Thus the proposed simulation model is convergent to Xiao’s model when the train is moving with a constant speed in a certain direction, overcoming the Xiao model’s limitations that only can be used for the simulation of constant speed. Taking into account the non-stationary characteristics due to ultra-high mobility, the thesis analyzes the stationarity region and coherence region, and discusses the time-varying channel model based on random birth-death process.Subsequently, the thesis focuses on the BEM(Basis Expansion Model) targeted at the specific characteristics of high-speed railway broadband wireless communication channel with both time and frequency selective fading. Based on several basic BEM models and the two modified models, the thesis proposes several improved methods, that is,(1) make a combination of two modified models,which can ensure the maximum spectrum frequency equal to the maximum Doppler frequency shift, and reduce the effects of Gibbs phenomenon;(2) partially overlap the two adjacent time frame in CE-BEM, which can effectively reduce the edge modeling errors, resulting better modeling performance than that of original CE-BEM and modified CE-BEM;(3) adopt segmentation modeling strategy by keeping each segment modeling order Q unchanged, which can achieve higher modeling accuracy than the original model. In the third improved method, when Q changes with the number of segments, the modeling performance becomes the best when the number of segments is two, and reduces as the number of segments increases due to the edge modeling error, but all are better than the CE-BEM and modified CE-BEM.Finally, the thesis simulates a simple OFDM wireless communication system, using LMMSE channel estimation. The BEM is used for the reconstruction of the channel. Simulation results show that the normalized mean square error of the first improved method (the combination of MCE-BEM and modified CE-BEM), is lower than the original corresponding model. In particular, under high mobility scenarios, the performance improvement is more obvious.