| Summary: | To combat the explosive growth of mobile data traffic, massive-multi-input multi-output (MIMO)-enabled wireless backhaul small-cell (SC) network (WBSN) has been investigated in recent years. One of the key challenges for the WBSN is to reduce the backhaul processing delay. However, a large portion of current research has mostly been focused on the decode-and-forward (DF) protocol that inherently contains a large transmission latency due to complicated decoding process at each SC base station (SBS). In this paper, we investigate the amplify-and-forward (AF)-based WBSN. Unlike the DF schemes, the AF-WBSN is more attractive due to the cost effectiveness and lower computational complexity. In frequency division duplex (FDD) systems, however, such advantages come at the expense of high channel estimation complexity, because the macro-cell base station (MBS) in the AF-WBSN is required to know the global channel state information, which causes a significant feedback delay as well as overhead. The channel estimation becomes even more challenging when it comes to the massive MIMO systems where the MBS is equipped with a large excess of antennas. To tackle the problem, we propose a novel transceiver design for the FDD massive-MIMO enabled AF-WBSN whose channel estimation and feedback complexity reduce to the level of its DF counterpart with a low latency by leveraging the antenna correlation at the MBS and the mean squared error decomposition properties at the SBSs. Finally, numerical results verify the efficiency of our proposed designs.
|
|---|
