Precoded OFDM With Spectrum-Aligned DFT for Extreme Doppler UWA Channels

Abstract

Orthogonal Frequency-Division Multiplexing (OFDM) systems without in-band pilots suffer from catastrophic Inter-Carrier Interference (ICI) and irreducible error floors when subjected to extreme Carrier Frequency Offset (CFO). While cross-domain waveforms like OTFS address high Doppler spreads, their detection complexity remains prohibitive for real-time edge hardware [4], [5]. In this letter, we propose a novel receiver paradigm termed Spectrum-Aligned Discrete Fourier Transform (SA-DFT). Rather than suppressing ICI with complex equalizers, SA-DFT re-aligns the DFT sampling grid to match the Doppler-shifted subcarrier frequencies, transforming the dominant CFO-induced ICI into a largely common phase rotation that can be blindly corrected. Synergized with Time-Division Duplexing (TDD) Regularized Zero-Forcing (RZF) precoding and a blind power-law estimator, the proposed holistic architecture guarantees reliable demodulation with zero in-band pilot overhead during data transmission. Extensive validations confirm that SA-DFT operates with <inline-formula> <tex-math notation="LaTeX">$\mathcal {O}(N^{2})$ </tex-math></inline-formula> complexity while delivering a 6.25% effective throughput gain and tolerating initial CFO errors up to <inline-formula> <tex-math notation="LaTeX">$\Delta \varepsilon =0.14$ </tex-math></inline-formula>. Applied to the Watermark BCH1 dataset, SA-DFT improves EVM from 45.3% down to 9.8% (<inline-formula> <tex-math notation="LaTeX">$4.6\times $ </tex-math></inline-formula>), demonstrating practical robustness.