Millimeter Wave Channel Estimation using Data-Aided DoA Estimation

Abstract

In this paper, we study channel estimation for multiuser millimeter wave (mmWave) systems. Since most conventional channel estimation schemes rely entirely on pilot signals, such methods cannot accurately estimate mmWave channels with low signal-to-noise ratio (SNR). To address this challenge, a direction of arrival (DoA)-assisted channel estimation scheme exploiting the sparsity of mmWave channels is developed. The proposed scheme utilizes data symbols as well as the pilot signals to estimate DoA of effective paths, and the estimated DoA information is used to suppress the noise given that the number of effective path directions is much smaller than the number of antennas. Furthermore, it is shown that sparse Bayesian learning (SBL) approach is suitable for DoA estimation in mmWave communications scenarios where signals from different angles are highly correlated as a result of multipath propagation, and a low complexity SBL algorithm for this setup is derived. By utilizing data symbols and adopting the SBL approach, the proposed scheme can identify the directions of effective path even in low SNR regime, enabling accurate estimation of mmWave channels.