Year: 2026
Type: preprint
Abstract: Orthogonal time frequency space (OTFS) modulation offers robust performance under high-mobility scenarios. Spread pilots distribute pilot energy across the delay-Doppler grid and are superimposed with data symbols, enabling full-grid resource utilization. However, pilot-data separation challenges the design of the real-time receiver pipeline: cross-ambiguity in channel sensing incurs high computational complexity, and pilot-data interference cancellation requires repeated processing through turbo iterations. Leveraging the chirp structure of OTFS spread pilots, we reduce cross-ambiguity to a single 2D IFFT with element-wise chirp operations, achieving over 7000x speedup compared to partially vectorized direct evaluation on GPUs at a frame size of (M, N)=(2053, 31). Our real-time GPU baseband processing pipeline for OTFS with superimposed spread pilots supports (M, N)=(2053, 31) with one turbo iteration within a 1.03 ms latency budget. We further show that larger frames increase processing gain, allowing lower pilot power and near-converged bit error rate (BER) after one turbo iteration.
Source: arXiv (Cornell University)
Institution: Duke University
Language: English
Topic: PAPR reduction in OFDM
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: green
Funders: National Science Foundation, Nvidia
OpenAlex ID: W7226000474
