Channel Estimation for Large Antenna Systems with 3GPP Channel Model

The advantage of large antenna systems relies on the acquisition of channel state information (CSI) at the base station (BS), which can be difficult especially for frequency-division-multiplexing (FDD) systems, as the pilot overhead increases linearly with the BS antenna number. In this paper, we exploit the channel sparsity to design a pilot transmission and channel estimation technique that significantly reduces the pilot overhead. Practical constraints and properties of physical propagations channels are considered to optimize our design. The proposed technique has two distinctive features. First, the channel estimation problem is decomposed into three sub-problems of delays, angles and (complex) amplitudes estimation of propagation paths, each of which has much lower complexity. Second, the angle estimation problem is turned into a cluster estimation problem, which has a much lower dimension and can be realized more efficiently and effectively. Simulation results show that for a large antenna system with 100 antennas, the proposed technique can reduce the pilot overhead to about 20% that of the conventional minimum-mean-square-error (MMSE) channel estimator, with only slight performance degradation.