Low Complexity 2D Off-Grid OTFS Channel Estimation in Fractional Delay-Doppler Scenarios

Abstract

In high-mobility wireless communication scenarios, orthogonal time-frequency space (OTFS) is a promising two-dimensional (2D) modulation scheme that can efficiently represent the delay-Doppler (DD) domain channel using only a few parameters. However, fractional delay and/or fractional Doppler reduce the channel’s sparsity, thus increasing the difficulty in channel estimation. Off-grid methods are previously utilized to estimate the fractional delay-Doppler channels, based on the expectation-maximization sparse Bayesian learning (EMSBL). However, these methods require matrix inversion in each iteration which results in significant computational complexity. To reduce the complexity, this paper proposes a low-complexity off-grid channel estimation scheme by employing a recently introduced efficient sparse Bayesian learning (ESBL). Additionally, a strategy for updating the off-grid parameters is adopted, further reducing the complexity of off-grid channel estimation. Simulation results demonstrate the effectiveness of the proposed method in comparison to existing channel estimation schemes in the doubly fractional scenarios, at a slight tolerable loss in channel estimation performance.