A SAGE Algorithm for Channel Estimation using Signal Eigenvectors for Direction-Scan Sounding

In this paper, the Space-Alternating Generalized Expectation-maximization (SAGE) approach is applied to estimating the parameters of multipath components (MPCs) in propagation channels with the observations being the signal eigenvectors (SEVs) calculated from channel impulse responses (CIRs) collected via direction-scan sounding (DSS), i.e. rotating a horn antenna in directions. The performance of such an SEV-based-SAGE (SEV-SAGE) algorithm is compared with the conventional Raw-Data-based SAGE (RD-SAGE) algorithm in terms of either root mean square estimation errors (RMSEEs) in single-path scenarios or the composite multipath component distances (MCDs) in multi-path scenarios. Furthermore, both algorithms are applied to the measured data for 13-17 GHz indoor propagation channels obtained via DSS. These results demonstrate the superiority of applying the SEV-SAGE algorithm over the RD-SAGE algorithm for the capability of estimating the overall channels with a fewer number of MPCs.