Channel Estimation for OTFS Waveform Using Kalman Filtering in LEO Satellite Communications

In low earth orbit (LEO) satellite communications, due to the effect of large delay and Doppler shift, it is difficult to obtain accurate channel estimation results in the classical time-frequency domain. Orthogonal time frequency space (OTFS), for its feature that only a signal can estimate entire-frame channel in the delay-Doppler (DD) domain, is regarded as a potential solution to this issue. In this paper, to catch the channel sparsity in the DD domain, two-dimensional (2D) orthogonal matching pursuit (OMP) algorithm is proposed. However, the presence of burst noise can lead to the incorrect judgment or loss of channel path in estimation, which can severely impact the performance of the 2D-OMP algorithm. Thus, the Kalman filtering is proposed, which can use a priori information to resist burst noise. In order to use the Kalman filtering, the mean square error (MSE) frameworks for OMP algorithm and prediction model are established, respectively. Simulation result shows that the proposed method can further improve the channel estimation accuracy with low computational complexity.