Research On NR System Enhancement Technology In High Speed Scenarios

The rapid development of high-speed rail makes travel more convenient,and provides new application and deployment scenarios for wireless communication.As one of the typical high-speed scenarios,high-speed rail scenario has not only the obvious Doppler effect and fast-varying characteristics,but also a series of characteristics of high-speed rail scenario layout.The commercial application of the 5th-Generation(5G)mobile communication technology has gradually expanded in the world.The Orthogonal Frequency Division Multiplexing system adopted by 5G NR(New Radio)is bound to face the challenges brought by this scenario.The destruction of orthogonality among sub-carriers due to the Doppler effect and the channel estimation error caused by time-varying channels have become the major technical challenges of NR high-speed scenarios.Firstly,this thesis studies the modeling of high-speed train scenarios,and then studies the enhancement technology based on the single frequency network model.In order to improve the channel estimation accuracy in high-speed scenarios,this thesis adopts a joint optimization scheme of frequency offset estimation and channel estimation.After studying three existing frequency offset estimation algorithms under the multi-path channel,this thesis proposes an algorithm of higher precision:linear fitting least square frequency offset estimation algorithm.Based on Wiener filtering,frequency offset estimation and channel estimation are combined to improve the channel estimation accuracy.Then the system performance of each algorithm is verified by simulation,and the performance differences of each algorithm and the performance advantages of the proposed algorithm are analyzed.Secondly,in order to solve the problem of selecting the best numerology for the NR system in high-speed scenarios,this thesis analyzes the defect of inadequate accuracy of the original channel model,and then extends the channel model.In case of the extended channel model,a simulation method based on sample points is designed.Performance evaluation of different numerologies under different Demodulation Reference Signal(DMRS)patterns is carried out,and the numerology configuration proposal for high-speed scenarios is given,which provides a reference for current mobile communication network deployment.Finally,frequency offset tracking is also a common method to enhance frequency offset estimation.This thesis studies the conventional frequency offset tracking scheme based on Kalman filtering.With the frequency offset variation characteristics in high-speed scenarios,a new frequency offset tracking algorithm is proposed based on the vary-by-phase characteristics of frequency offset.This thesis verifies its performance advantages by simulation:it can not only track the fast-changing frequency offset more accurately,but also effectively reduce the influence of residual error.