Research On AoA Estimation And Positioning Algorithms For 5G System

As location information becomes vitally important,positioning has been a highly desirable feature of 5G system which enables a huge amount of location-based applications and services.New scenarios and services have more strict requirements on positioning technology precision and forms.Millimeter wave(mmWave)is the promising technology for 5G system having quasi-optical and sparse multi-path propagation properties.mm Wave is always deployed with massive MIMO for that they can complement each other to overcome some drawbacks offering better performance for AoA based positioning.Thus,this research pays attention to indoor AoA estimation and positioning schemes.The paper is summarized as follows:1.Propose AoA estimation methods for 5G indoor low Signal to Interference and Noise(SINR)scenarios.First of all,using the uplink Sound Reference Signal(SRS)as the positioning reference signal,combined the Poisson point process theory we analyze and simulate to validate that 5G indoor office is a typical low SINR scene.Based on this,we propose the Toeplitz adjustment and eigenvalue noise suppression methods,and design a customized peak search strategy to improve the traditional multiple signal separation algorithm(MUSIC).These schemes can effectively improve the angle estimation performance under low SNR,and the eigenvalue suppression method has better robustness than the Toeplitz method.2.Design a indoor positioning and track tracing algorithm based on mm Wave multi-path information and unscented Kalman filter.User equipment(UE)motion feature,mm Wave line-of-sight(LoS)and first order reflection paths' AoA-ToA are fused for indoor positioning.The problem of insufficient anchor nodes can be solved by introducing additional mirror base stations without adding basic physical devices.We analyze the technical requirements of Internet of things application for real-time positioning and track tracing of indoor UEs,and build 5G indoor positioning system model.Furthermore,a modified multipath unscented Kalman filter(UKF)is proposed to weaken the nonlinear relationship between user position and measurement variables.Historical information combining with the measurement of the angle of arrival(AoA)and the time of arrival(ToA)of multipath signals,the distribution parameters for the noise error of the measured value are estimated and corrected.Finally,the corrected result is sending back to the user's position estimation value,so as to eliminate part of the positioning error and improve the positioning accuracy.