Prediction of Time-Varying Multi-User MIMO Channels Based on DOA Estimation Using Compressed Sensing

Abstract

In multi-user multiple-input multiple-output (MIMO) systems, a base station requires downlink channel state information for each user to suppress interferences. In time-varying environments, however, channels at the actual transmission time differ from the estimated ones because we have delay. This causes interference and deteriorates the packet error rate performance. To overcome this problem, we previously proposed a channel prediction method based on direction of arrival (DOA) estimation of multipath components obtained with a compressed sensing technique. This method can predict channels with high accuracy when we can assume that user equipment alone moves and the surrounding scatterers are stationary. In real environments, however, this assumption does not hold, and this method cannot be applied. In this paper, we propose a prediction method using two-step compressed sensing without the above assumption. In the first step, we estimate the DOAs of multipath signals. In the second step, we estimate the Doppler frequencies and their complex amplitudes using the estimated DOA information. From these values, we predict the MIMO channels at the actual transmission time. We discuss the performance of this method in this paper.