Channel equalization using classification methods

Abstract

This thesis has been done in Digital and Computer Systems Laboratory at Tampere University of Technology. This work was part of technology programs called "Adaptive and Intelligent Systems Applications" and "User-Oriented Information Technology", set up by the National Technology Agency. In this work we concentrated especially on digital mobile radio systems, which suffer from intersymbol interface due to multipath propagation. We investigated two classification methods and compared them to some of the traditional channel equalization techniques. We also investigated some modifications of the basic algorithms. In addition to the original major voting rule we evaluated a distance-weighting alternative /Kir10