Optimization of linear MIMO precoding assuming MMSE-based turbo equalization

In this paper, we focus on the optimization of a linear multiple-input multiple-output (MIMO) precoder assuming an outer forward error correction code and an iterative minimum mean square error (MMSE)-based interference cancellation (turbo equalization) at the receiver side. Given perfect channel state information at both sides of the communication, we propose a novel precoder that is specifically designed to use with turbo equalization. In contrast to the conventional precoders that maximize the mutual information (MI) between finite alphabet input and the corresponding output over the precoded MIMO channel, the proposed precoder aims to maximize the MI between the finite alphabet input and the corresponding output of the equalizer. The precoder is targeted for applications that require low complexity, where forward error correction (FEC) codes with low-complex encoding and decoding structures are used. Simulation results show the error-rate performance gain of the resulting precoder compared to two other reference precoders presented in the literature, which are derived from the maximization of the precoded MIMO channel MI.