Unfolded methods for channel estimation denoising

Thumbnail Image

Files

master_Wong_Wei_Kang_2025.pdf (1.02 MB)

URL

Journal Title

Journal ISSN

Volume Title

School of Electrical Engineering | Master's thesis
Unless otherwise stated, all rights belong to the author. You may download, display and print this publication for Your own personal use. Commercial use is prohibited.

Authors

Date

2025-11-22

Department

Major/Subject

Speech and Language Technology

Mcode

Degree programme

Master's Programme in Computer, Communication and Information Sciences
Tieto-, tietoliikenne- ja informaatiotekniikan maisteriohjelma
Magisterprogrammet i data-, informations- och kommunikationsteknik

Language

en

Pages

33

Series

Abstract

In wireless communications, accurate channel estimation is critical for reliable signal recovery, particularly in multiple-input multiple-output (MIMO) systems. Pilot observations first yield a least squares (LS) estimate of the channel, which provides an unbiased but noisy observation of the true channel. To refine this estimate, traditional methods such as hard windowing apply fixed linear filters to denoise the LS output, but their non-adaptive nature can suppress valid channel components and degrade performance when noise levels or channel statistics vary. This thesis investigates the use of algorithm unfolding for improving the denoising of the LS channel estimate. We focus on LIDIA (Learned Iterative Denoising Algorithm), an unfolded neural network that mimics the structure of classical iterative denoisers while incorporating learnable components. The model combines the interpretability of traditional iterative algorithms with the adaptability of deep learning. Experimental results show a 4.43 dB improvement over baseline denoisers, while requiring up to 5.4× fewer parameters than transformers. This highlights the potential of unfolded denoising networks as lightweight alternatives to heavy deep-learning estimators such as transformers.

Description

Supervisor

Alku, Paavo

Thesis advisor

Sheriff, Mohammed Rayyan

Keywords

LIDIA, channel estimation, denoising, unfolding, unrolling, lightweight

Other note

Citation

Permanent link to this item

https://urn.fi/URN:NBN:fi:aalto-202512169318

Collections

[dipl] Sähkötekniikan korkeakoulu / ELEC