Optimizing ZF Precoders for MIMO Broadcast Systems

In this paper, we develop optimization techniques for linear zero-forcing (ZF) multiple-input multiple-output (MIMO) broadcast (BC) precoders. A computationally efficient technique based on QR-update algorithms for implementing this optimization is then proposed. The application of the proposed ZF precoders and decoders to both large and small (in terms of the number of users) systems is considered. It is also shown that the proposed optimization technique is robust, and performs very well with nonlinear Tomlinson--Harashima precoding. Both coded and uncoded performances are considered.