Minimum Mean-Squared Error Multi-Branch Decision Feedback Detection for MIMO Systems

In this paper we propose a novel decision feedback (DF) detection strategy for multi-input multi-output (MIMO) spatial multiplexing systems based on multiple processing branches. The proposed detection structure employs multiple cancellation branches with appropriate transformations that modify the initial design. The novel structures for detection exploit different patterns and orderings for the modification of the original feedforward and feedback filters design. We present minimum mean-squared error (MMSE) estimators for the multi-branch structure and the design of MIMO receivers and show that the proposed MMSE design does not require a significant additional complexity. The proposed multi-branch MMSE DF detectors are compared with several existing detectors in the literature via computer simulations and are shown to close the gap between MMSE detectors and the optimal maximum likelihood detector.