EUVIMEDEUVIMED entdecken
Crossref

Online Structured Channel Estimation for Time-Varying mmWave MIMO-OFDM via Bayesian Tucker Modeling

Tianyu Jiang

IEEE Signal Processing Letters · 2026 · 33 · 2525-2529

DOI: 10.1109/lsp.2026.3701618 ↗

Kurz gefasst

Fachartikel in der Zeitschrift IEEE Signal Processing Letters, 2026. Verfasst von Tianyu Jiang. Die Arbeit nennt 12 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

This letter proposes an online channel estimation framework for time-varying millimeter-wave (mmWave) MIMO-OFDM systems with hybrid beamforming.The received training data in each frame are organized into a third-order tensor over the combiner, subcarrier, and pilot dimensions, and a Bayesian Tucker model is used to obtain its structured estimate, with a spike-and-slab (SS) prior imposed on the core tensor. The channel matrix of each subcarrier is then recovered from the reconstructed tensor through the known training relationship. To enable online operation in high-mobility scenarios, we develop a streaming variational inference algorithm with moment matching and subspace tracking. Simulation results show that the proposed method achieves higher estimation accuracy and stronger robustness, especially in fast time-varying environments.

Zitieren

APA

Jiang, T. (2026). Online Structured Channel Estimation for Time-Varying mmWave MIMO-OFDM via Bayesian Tucker Modeling. IEEE Signal Processing Letters, 33, 2525-2529. https://doi.org/10.1109/lsp.2026.3701618

Vancouver

Jiang T. Online Structured Channel Estimation for Time-Varying mmWave MIMO-OFDM via Bayesian Tucker Modeling. IEEE Signal Processing Letters. 2026;33:2525-2529. doi:10.1109/lsp.2026.3701618

BibTeX
@article{Jiang2026,
  author = {Jiang, Tianyu},
  title = {Online Structured Channel Estimation for Time-Varying mmWave MIMO-OFDM via Bayesian Tucker Modeling},
  journal = {IEEE Signal Processing Letters},
  year = {2026},
  volume = {33},
  pages = {2525--2529},
  doi = {10.1109/lsp.2026.3701618}
}
Kostenlos starten →