Impulse-Pilot-Aided OMP Channel Estimation for SEFDM-Based OTFS Systems

Orthogonal time frequency space (OTFS) modulation offers robust performance over doubly selective channels, yet its spectral efficiency is constrained by the conventional orthogonal subcarrier spacing. To address this limitation, spectrally efficient frequency division multiplexing (SEFDM)-based OTFS intentionally compresses the subcarrier spacing in time-frequency domain to enhance spectral efficiency while retaining the delay-Doppler characteristics of OTFS, at the cost of inherent inter-carrier interference. In this paper, we propose an OMP-based sparse channel estimation framework for SEFDM-based OTFS using an impulse-pilot block at the transmitter. We obtain the corresponding dictionary by deriving the delay-Doppler (DD) domain input-output relationship of the SEFDM-based OTFS system, and address the colored-noise issue induced by subcarrier compression by adopting generalized least squares. Moreover, we derive both the Cramér-Rao lower bound (CRLB) with unknown support and the oracle CRLB with known support as theoretical performance benchmarks for the proposed channel estimation algorithm. Simulation results demonstrate that the proposed algorithm closely approaches the oracle CRLB and exhibits negligible performance variation across different packing factors.