Research On Beam Tracking Technology In Millimeter Wave System

Along with the application research of millimeter wave communication technology in 5G mobile communication,the challenges brought by it have also attracted the attention of researchers.The robustness of mmWave system is a major challenge because mmWave is usually transmitted in a directional manner and is oriented to mobile scenarios.In order to ensure high quality communication performance,we need to track the millimeter wave beam in real time.Firstly,the development and current situation of mmWave beam direction estimation technology are summarized,the signal model and channel model are introduced,and the basic mathematical model of beam estimation is given.Then the beamforming technique and the classical spatial spectral estimation algorithm are introduced.Then,two beam estimation methods,gradient descent method and Newton search method,are compared by simulation,and their convergence is analyzed.Secondly,the thesis studies the adaptive beam tracking technology based on extended kalman filter,and completes an adaptive estimation of the Angle of the estimated signal source in the case of insufficient pilot signal,so as to improve the performance of beam tracking.In this thesis,the extended kalman filter tracking technique is applied to the millimeter wave two-dimensional planar array,and the joint tracking of phase noise and Angle is considered.Then,the performance of extended kalman filter in line array and plane array is simulated.Thirdly,the expectation maximization algorithm and the improved spatial alternating generalized EM algorithm are applied to the Angle estimation problem of millimeter wave.These two algorithms avoid the multi-dimensional search of maximum likelihood algorithm parameter estimation,simplify the computational complexity,and improve the accuracy.Then,the performance and convergence of the two algorithms are compared by simulation.Finally,considering the large bandwidth of mmWave system and the decrease of efficiency caused by simulated beam scanning,the blind beam tracking problem is studied.Markov chain monte carlo algorithm is applied to joint beam tracking and signal detection in mmWave system.Simulation results show that this method can achieve better tracking performance and symbol detection performance.