Dirty paper coding based optimal MIMO relay communications

In this paper, we address the optimal source and relay matrices design issue for a multiple-input multiple-output (MIMO) relay network using the dirty paper coding (DPC) scheme at the source node. The aim is to minimize the mean-squared error (MSE) of the signal waveform estimation at the destination. Using the property of uplink-downlink duality, the original DPC-based MIMO relay system is first converted to a dual system with a decision feedback equalizer (DFE) at the destination. Then we jointly optimize the source and relay matrices of the dual system. Finally the optimal source and relay matrices of the DPC-based system are obtained by exploiting the link between the source, relay, and destination matrices of the original and dual MIMO relay systems. Simulation results demonstrate that the proposed DPC-based MIMO relay system performs much better than the existing linear minimal MSE (MMSE)-based relaying approach in terms of bit-error-rate.