Utilizing channel mean information with MMSE detection for LDC in MIMO systems

In this work, we study a MIMO system with a transmitter using a linear dispersion code (LDC) and a linear MMSE detector at the receiver in a Ricean flat-fading environment. The focus of our work is the analysis of the optimal transmit strategy using different types of LDC assuming that the transmitter knows only the mean channel matrix. On the one hand we consider spatial multiplexing (SM) schemes achieving high data rates, however sacrificing diversity. On the other hand, we have schemes achieving full diversity like quasi-orthogonal space-time block codes (QSTBC) or orthogonal space-time block codes (OSTBC). Depending on the LDC in use, the optimization problem is either convex or non-convex. For both these classes of LDC we derive the properties of the average normalized MSE and analyze the impact of the mean component on the MSE, the optimal transmit strategy and the optimal power allocations.