Linear Precoding for MIMO Systems

Broadband transmission with antenna diversity at both the transmitter and receiver sides offers several degrees of freedom to design modems. They can be exploited to increase the resilience or increase the throughput of the communication link. The wireless channel effect can be mapped conveniently in a multiinput multi-output (MIMO) model, which allows the derivization and analysis of linear code designs using algebraic tools. We refer to these techniques as linear precoding methods. This chapter focuses on three aspects related to the design of such precoders: (1) the optimal design under average and peak power type of constraints; (2) the performance analysis in the presence of random fading; and (3) the performance of generalized training and semiblind techniques.