Achievable rates for generalized spatial Tomlinson-Harashima precoding in MIMO systems

In this paper, we study spatial Tomlinson-Harashima precoding (STHP). We propose a generalization of the original structure of STHP which enables different transmission power per antenna. The main advantage of the new scheme is that it has a scalable structure, because, despite the variable power per transmitter (even if some transmitters are switched off), the precoding structure and its properties are preserved. Next, we formulate the maximum achievable rates problem for GSTHP and solve it numerically. In addition, we show that the problem can be quasi-optimally solved in a closed form in a limit case, or if some approximations are done. The solution is found to be an antenna-selection algorithm with uniform power allocation among the active transmitters. We finally present some simulation results and conclusions.