Sparse Bayesian Learning of Delay-Doppler Channel for OTFS System

Abstract

Orthogonal time frequency space (OTFS) modulation has superior performance than traditional orthogonal frequency division multiplexing (OFDM) in fast time-varying scenarios. However, due to the effect of Doppler shift, higher pilot overhead and pilot power are required to estimate the channel state information. According to the sparseness of the channel in the delay-Doppler domain, this letter proposes a new pilot pattern and a sparse Bayesian learning (SBL)-based channel estimation algorithm. There is no guard pilot in the pilot pattern, and the pilot has the same power as the data. Based on the new pilot pattern, we first convert the channel estimation problem to a sparse signal recovery problem. Then, we introduce a sparse Bayesian learning framework and construct a sparse signal prior model as a hierarchical Laplace prior. Finally, the expected maximum (EM) algorithm is used to update the parameters in the prior model. Numerical simulation highlights the superiority of the proposed algorithm in terms of pilot overhead, pilot power consumption, and anti-noise interference.