Fast Subspace Modal Tracking for MIMO-OFDM Channels

Multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) wireless channels have slow-varying space and fast-varying time characteristics through modal analysis. In this paper, we investigate subspace modal tracking for estimation of the channel finite impulse responses. Since known channel structures facilitate space-time decouple, the common space mode is estimated as a whole for all the shared subcarriers. A recursive fast subspace approximation tracking algorithm combined with a least square (LS) estimator is applied to track space-time modes under non-stationary noise circumstances. Simulation demonstrates the hybrid approach has a good mean-square error performance. Up to 3dB gain can be obtained for certain SNRs over the LS estimator.