English

DMD Prediction of MIMO Channel Using Tucker Decomposition

Information Theory 2026-03-17 v1 Signal Processing math.IT

Abstract

Accurate channel state information (CSI) prediction is crucial for next-generation multiple-input multiple-output (MIMO) communication systems. Classical prediction methods often become inefficient for high-dimensional and rapidly time-varying channels. To improve prediction efficiency, it is essential to exploit the inherent low-rank tensor structure of the MIMO channel. Motivated by this observation, we propose a dynamic mode decomposition (DMD)-based prediction framework operating on the low-dimensional core tensors obtained via a Tucker decomposition. The proposed method predicts reduced-order channel cores, significantly lowering computational complexity. Simulation results demonstrate that the proposed approach preserves the dominant channel dynamics and achieves high prediction accuracy.

Keywords

Cite

@article{arxiv.2603.15468,
  title  = {DMD Prediction of MIMO Channel Using Tucker Decomposition},
  author = {Irina Kopnina and Dmitry Artemasov and Sergey Matveev},
  journal= {arXiv preprint arXiv:2603.15468},
  year   = {2026}
}

Comments

This work has been submitted to the IEEE for possible publication

R2 v1 2026-07-01T11:22:34.478Z