OPTIMIZATION IN DESIGN OF PRECODER AND DECODER FOR COORDINATED ACCESS POINTS SYSTEMS USING MMSE CRITERION IN MEASURED CHANNELS
Author
Matallana Espinar, Carlos
Term
4. term
Publication year
2011
Pages
96
Abstract
Denne afhandling udvikler en iterativ algoritme til at designe lineære sende- og modtage-beamformingfiltre, der maksimerer den samlede datahastighed (sum-rate) i et kooperativt MIMO-system. MIMO (multiple-input multiple-output) bruger flere antenner til at sende og modtage, og i kooperative systemer koordinerer flere sendere deres signaler. Beamformingfiltre former signalerne, så de bedst når frem til de ønskede modtagere. Kanalinformation (Channel State Information, CSI) for alle forbindelser antages kendt ved senderne. Metoden bygger på nyere arbejde, der forbinder Weighted Minimum Mean Square Error (WMMSE) med Weighted Sum Rate (WSR) i MIMO-broadcastkanaler. Denne forbindelse gør det muligt at beregne sende-filteret med lav beregningskompleksitet, og numeriske resultater viser, at algoritmen opnår sin ydeevne efter få iterationer. Projektet undersøger også, hvordan algoritmen fungerer på virkelige måledata indsamlet i Aalborg, Danmark. Vi viser, hvordan data kan forbehandles med signalbehandling for at sammenligne deres adfærd med teoretiske kanalmodeller, og vi afprøver flere bearbejdningsmåder for at se, hvordan de påvirker brugernes opnåelige hastighed. Resultaterne viser, at virkelige kanaler ikke når den bedste rateydeevne på grund af uundgåelige uregelmæssigheder; den største negative effekt kommer fra ubalance mellem grenenes effektniveauer (branch power ratio).
This thesis develops an iterative algorithm to design linear transmit and receive beamforming filters that maximize the total data rate (sum-rate) in a cooperative MIMO system. MIMO (multiple-input multiple-output) uses multiple antennas to send and receive, and in cooperative systems several transmitters coordinate their signals. Beamforming filters shape signals so they reach intended users more effectively. Channel State Information (CSI) for all links is assumed to be known at the transmitters. The method builds on recent work linking Weighted Minimum Mean Square Error (WMMSE) to Weighted Sum Rate (WSR) in MIMO broadcast channels. This connection enables a low-complexity solution for the transmit filter design, and numerical results show the algorithm reaches its performance in few iterations. The project also examines how the algorithm behaves with real measurement data collected in Aalborg, Denmark. We demonstrate signal-processing steps to prepare real data for comparison with theoretical channel models and test several processing approaches to assess their impact on users’ achievable rates. Findings indicate that real channels do not achieve the best rate performance due to imperfections; the largest detrimental effect is the imbalance in branch power ratios.
[This abstract was generated with the help of AI]
Documents