Research On Millimeter Wave Beam Tracking Algorithm

Millimeter wave communication technology is one of the key technologies for 5G high-frequency communication.In order to effectively compensate for the huge path loss in the propagation process,the millimeter wave communication system usually uses beamforming(BF)to focus the energy of the wireless signal,thereby forming a directional narrow beam.But how to keep the beams between the base station and the terminal aligned,especially when the user is in a mobile condition to quickly and accurately perform beam pairing is a technical difficulty.To obtain a directional narrow beam with high gain,the transceiver needs to obtain accurate Channel State Information(CSI).Traditional solutions generally use channel estimation technology,but this solution will cause huge training overhead in dynamic scenarios.Therefore,this thesis mainly studies low-complexity beam tracking technology suitable for millimeter-wave communication systems based on two uniform linear/rectangular transmission models.First,a low-complexity beam tracking algorithm based on a Uniform Linear Array model is studied.The algorithm takes the second-order extended Kalman filter as the core and combines a beam switching scheme based on threshold decision to reduce the estimation error while ensuring low complexity.Simulation results show that when the channel angle parameter AOA/AOD(Angles of Arrival/Angles of Departure)changes in the speed range of 0.5?0.7 degrees/slot,this algorithm can reduce the estimation error by up to about 30% compared with similar algorithms,the computational complexity is still kept low.Then,in order to solve the problem that large arrays are sensitive to AOA/AOD,a robust two-stage beam tracking algorithm is proposed.This algorithm improves the structure of Auxiliary Beam Pair(ABP)to make it suitable for beam tracking framework,and uses Extended Kalman Filter(EKF)beam tracking as a first-step algorithm to realize the composition of initial ABP.This method cleverly implements a low complexity beam tracking algorithm with AOA/AOD correction capability.Simulation results show that the algorithm effectively reduces the sensitivity of large arrays to AOA/AOD changes,so that the trajectory of AOA/AOD estimates can closely follow the trajectory of real values,which significantly improves the robustness of the same type of algorithm.Finally,a beam tracking algorithm based on a uniform rectangular array model is studied.Firstly,the solution ideas of the existing Matrix Factorization(MF)calibration algorithm are analyzed,and the problems of the large amount of calculation and the idealization of cost function selection caused by the element-level calibration of the channel matrix by this algorithm are pointed out.Therefore,a suboptimal calibration algorithm based on the Minimum Squared Error(MSE)criterion is studied,and an initial estimation is attempted using Unscented Kalman Filter(UKF)beam tracking.Simulation results show that the UKF beam tracking algorithm can effectively reduce the number of iterations of the calibration algorithm,and the estimated performance of the MSE sub-optimal calibration algorithm in a slow time-varying channel environment can approximate the optimal MF calibration algorithm,and it has a smaller computation.