On subspace channel estimation in multipath SIMO and MIMO channels

Abstract

Earlier studies showed that iterative processing at receiver may significantly improve performance in wireless communications. In particular, performance of space-time block codes (STBC) extended for frequency selective channels may be notably improved by iterative channel estimation (ICE). On the other hand, the ICE at least doubles the receiver complexity that limits its implementation in handsets. In this paper we address non-iterative blind subspace estimation based on the second order statistics (SOS), which is simpler than the iterative receiver. Accuracy and robustness of the SOS (semi-)blind estimators with respect to different channel profiles are evaluated and compared to the ICE. Based on sensitivity analysis appropriate parameters for (semi-)blind estimators are selected. It is found that the SOS semi-blind subspace methods may approach accuracy (in terms of MSE) of the ICE. © Springer-Verlag 2004.