Channel Estimation for Large Antenna Systems

The advantage of large antenna systems relies on the acquisition of the channel state information at the base station, which is difficult for frequency-division-multiplexing systems, as the pilot overhead increases linearly with the antenna number. In this paper, we utilize the sparse and clustering characteristics of multipath channels to solve this problem. We decompose channel estimation into three low-complexity low-overhead parts: the estimation of delay, active bins and (complex) amplitudes of clusters of propagation paths. In each part, efficient pilot transmission and channel estimation are designed to leverage the features of the parameter to be estimated. Compared with the conventional minimum-mean-square-error channel estimator, the proposed technique significantly reduces the pilot overhead. Compared with compressive sensing based techniques that exploit only the channel sparsity but not the clustering distribution of paths, it achieves better performance with a much lower complexity.