Rapidly Varying Sparse Channel Tracking with Hybrid Kalman-OMP Algorithm

Abstract

It is expected from future communication standards that channel estimation algorithms should be able to operate over very fast varying frequency selective channel models. Therefore, in this study autoregressive (AR) modeled fast varying channel has been considered and tracked with Kalman filter over one orthogonal frequency division multiplexing (OFDM) symbol. Channel sparsity is exploited which decreases the complexity requirements of the Kalman algorithm. Since Kalman filter is not directly applicable to sparse channels, orthogonal matching pursuit (OMP) algorithm is modified for AR modeled sparse signal estimation. Also, by using windows, sparsity detection errors have been decreased. The simulation results showed that sparse fast varying channel can be tracked with the proposed hybrid Kalman-OMP algorithm and windowing method offers improved MSE results.