Student-T Prior Sparse Bayesian Learning for Improved Channel Estimation in OTFS Systems

Orthogonal Time-Frequency-Space (OTFS) modulation can effectively suppress the effects of Doppler shift in high-speed mobile scenarios. At the same time, the accuracy of OTFS channel estimation is an important factor that affects the performance of OTFS. In this paper, we propose a Sparse Bayesian Learning (SBL) algorithm to quickly and accurately estimate the OTFS channel by combining the pilot pattern and the sparsity of the OTFS channel. First, we propose a new pilot pattern to prevent the contamination of information symbols on pilot symbols. Since the pilot pattern uses only partial guard symbols, it can also improve the spectral efficiency. Then, we propose the Student-T prior SBL (STSBL) algorithm to improve the speed and accuracy of OTFS channel estimation by exploiting the sparsity of the OTFS channel. Simulation results show that the normalized mean squared error (NMSE), bit error rate (BER), and throughput of the proposed scheme outperform the benchmark schemes.