Mobile Millimeter Wave Channel Tracking: A Bayesian Beamforming Framework Against DOA Uncertainty.

Yan Yang,Shuping Dang,Miaowen Wen,Shahid Mumtaz,Mohsen Guizani

DOI: https://doi.org/10.1109/tcomm.2020.3026377

2019-01-01

Abstract:A Bayesian approach for joint beamforming and tracking is presented, which is robust to uncertain direction-of-arrival (DOA) estimation in millimeter wave ( mmWave) multiple input multiple output (MIMO) systems. The uncertain or completely unknown DOA is modeled as a discrete random variable with a priori distribution defined over a set of candidate DOAs, which describes the level of uncertainty. The estimation problem of DOA is formulated as a weighted sum of previously observed DOA values, where the weights are chosen according to a posteriori probability density function (pdf) of the DOA. In particular, we present a motion trajectory-based a priori probability approximation method, which implies a high probability to perform a directional estimate within a specific spatial region. We demonstrate that the proposed approach is robust to DOA uncertainty, and the beam tracking problem can be addressed by incorporating the Bayesian approach with an expectationmaximization (EM) algorithm. Simulation results validate the theoretical analysis and demonstrate the effectiveness of the proposed solution.