MmWave Beam Tracking With Spatial Information Based on Extended Kalman Filter

Abstract

Beam tracking is essential to compensate for beam misalignment and reduce the training overhead in millimeter wave (mmWave) systems. In high mobility scenarios, beam tracking can be implemented with the spatial information of the user equipment (UE) to avoid huge training overhead. However, there are estimation errors of spatial information in practical systems, which may lead to beam misalignment and performance degradation. In this letter, we propose a beam tracking algorithm to improve the tracking accuracy by combining spatial information and beam training. We first propose a novel channel evolution model based on the spatial information under the line-of-sight (LOS) condition. Then we consider the estimation error and modify the model accordingly. Based on the model, the extended Kalman filter (EKF) algorithm is utilized to track the beams at both ends. We also propose a novel training beam design to mitigate the non-linearity of measurement function and improve the tracking accuracy. Finally, simulation results validate the performance of the proposed algorithm.