Model-Driven Channel Estimation for OFDM Systems Based on Image Super-Resolution Network

TL;DR

This paper introduces a hybrid model-driven approach for OFDM channel estimation that combines traditional LS estimation with a CNN-based super-resolution network, achieving higher accuracy and spectrum efficiency.

Contribution

It presents a novel integration of image super-resolution techniques into channel estimation, bridging traditional methods with deep learning for improved performance.

Findings

Outperforms LMMSE estimation significantly

Potential for spectrum saving demonstrated

Analyzes impact of network structure and SNR mismatch

Abstract

In this paper, we propose a model-driven channel estimation method utilizing a convolutional neural network (CNN) derived from image super-resolution (SR). Instead of completely abandoning traditional communication modules as data-driven approaches always do, we first obtain rough channel matrix by the simplest least square (LS) channel estimation, and then use a CNN-based SR network to learn the subtle connections within channel matrix for accuracy improvement. The results show that our proposed method significantly outperforms the linear minimum mean squared error (LMMSE) estimation and has potential in spectrum saving. The impacts of network structure and SNR mismatch problem are also investigated in this paper.