Robust Update Algorithms for Zero-Forcing Detection in Uplink Large-Scale MIMO Systems

Abstract

A zero-forcing (ZF) detector is proposed for the uplink channel of large-scale multiple-input-multiple-output systems based on the URV decomposition of the channel matrix. Two efficient algorithms are also proposed for updating and downdating the channel decomposition matrices upon a user’s arrival or exit of the cell, respectively, which avoids re-calculating the URV decomposition. Specifically, the decomposition is updated through the Gram–Schmidt procedure when a user equipment (UE) joins the cell, and is downdated through a set of Givens rotations when a UE leaves the cell. The computational complexity of the proposed detector (referred to as the URV-ZF detector) is evaluated and compared with that of existing methods. It is shown through numerical studies that the URV-ZF detector achieves a lower bit error rate than existing schemes with a slight increase in the computational complexity.