Adaptive algorithms for group interference suppression for differential spatial multiplexing

Recently, we have proposed a novel differential space-time (DST) architecture (Cheung, S.K. and Schober, R., IEEE Wireless Commun. and Networking Conf., WCNC, 2004) which is much more robust against carrier phase variations than corresponding coherent ST schemes (Tarokh, V. and Lo, T.K.Y., 1998; Tarokh et al., 1999; Su, H.-J. and Geraniotis, E., 2002). To achieve a linear decoding complexity, we employ group interference suppression (GIS) and subsequent decision-feedback differential detection (DF-DD) at the receiver. The optimization of the GIS filters requires a generalized eigendecomposition of two correlation matrices. In order to reduce the computational complexity, we devise three efficient adaptive algorithms for generalized eigendecomposition. These algorithms are shown to have excellent convergence properties, and preserve the robustness of the proposed DST scheme against carrier phase variations.