A Selective Feedback Detection Algorithm Based on FDE-NP for MIMO Systems

Compared with the conventional FDE scheme, the FDE-NP structure which consists of a linear frequency domain equalizer (FDE) and a time domain noise predictor (NP) has lower complexity and achieves better performance and complexity trade-off. In this paper, we propose a selective detection algorithm for FDE-NP model in MIMO system, which we refer to as selective feedback detection (FDE-NP-SEL) algorithm. By evaluating the average MMSEs during the data transmission for all the possible feedback modes, the new detection algorithm chooses the optimal one to design time-domain NPs. In particular, to simplify the design of the feedback filters, we consider an inverted ordered detection (FDE-NP-INV) algorithm, in which decisions are fed back to predict and cancel their pre-cursor inter-symbol interference (ISI) effects to the previous signals. Simulation results show that the new proposed FDE-NP-INV scheme achieves significant performance improvement over the conventional FDE-NP scheme with the same complexity. Furthermore, we propose the hybrid ordered detection (FDE-NP-HYB) algorithm by combining the FDE-NP-INV algorithm with the conventional FDE-NP scheme . It is shown that the proposed FDE-NP-HYB can further improve the performance.