Institutional Repository
Technical University of Crete
Technical University of Crete
EN
|
EL
| URI: | http://purl.tuc.gr/dl/dias/AA6B1F27-842D-4B34-8649-B54AA6724EA3 | ||
| Year | 2025 | ||
| Type of Item | Diploma Work | ||
| License |
|
||
| Bibliographic Citation | Georgia-Nefeli Xynogala, "5G Wireless communication systems: Channel sparsity in the time and the angular domains", Diploma Work, School of Electrical and Computer Engineering, Technical University of Crete, Chania, Greece, 2025 https://doi.org/10.26233/heallink.tuc.104079 | ||
| Appears in Collections |
In this Diploma Thesis, we examine two sparsity-driven downlink channel estimation algorithms for Frequency Division Multiplexing (FDD) massive MIMO systems in 5G networks. Although the Base Station's (BS) massive antenna arrays offer significant improvements in spectral efficiency, capacity, and reliability, they also significantly increase the overhead of downlink Channel State Information (CSI) acquisition. To address this challenge, we explore algorithms that exploit channel sparsity in both the time and angular domains to enable more efficient and accurate CSI estimation.First, we present the fundamentals of compressive sensing and block sparse modelling in the time domain, and detail the algorithm of Cirik et al. (2016), which we reformulate as a convex optimization problem solved via CVX to retain the channel's block structure with improved convergence and complexity. Using the Orthogonal Matching Pursuit-based method of Alkhateeb et al. (2016), we then revisit angular-domain sparsity in millimetre-wave and hybrid MIMO channels and present a custom matching-pursuit variant that is adapted to realistic hybrid architectures and dictionary constraints.We apply both techniques in a MATLAB simulation framework using standardized 5G channel models, assessing computational load, convergence speed, and normalized mean-squared error (NMSE) at different Signal-to-Noise Ratio (SNR) and sparsity levels. The comparative analysis demonstrates how each approach strikes a balance between computational efficiency and estimation accuracy in various deployment scenarios. We conclude by outlining future research directions on adaptive dictionary learning and pilot-matrix optimization, as well as recommendations for choosing and fine-tuning these estimators in practical systems.
Copyright © DIAS 2013
