Complex-Valued GNN Based Detector for OTFS Signal under Imperfect Channel Information

Yue, Yang and Li, Zan and Shi, Jia and Ni, Qiang (2026) Complex-Valued GNN Based Detector for OTFS Signal under Imperfect Channel Information. IEEE Internet of Things Journal. ISSN 2327-4662

Text (Author final version)
Author_final_version.pdf - Accepted Version
Available under License Creative Commons Attribution-NonCommercial.
Download (4MB)

Abstract

In recent years, orthogonal time frequency space (OTFS) technique has garnered substantial academic attention as a promising solution for ensuring robust and reliable communication in high-mobility wireless communication environments. In this paper, we present a complex-valued graph neural network (CV-GNN) aided signal detection scheme for OTFS modulation, which can mitigate the channel spreading caused by fractional Doppler shifts. To mitigate inter-carrier interference (ICI) and inter-symbol interference (ISI) induced by fractional Doppler shifts and imperfect channel state information, the proposed detector is able to process the received OTFS signal in the complex plane to acquire the complete phase information of effective channel. Simulation results demonstrate that the proposed method can outperform other state-of-the-art schemes by 1∼4 dB in terms of reliability performance.