Modeling and Mitigation of Time- and Frequency-Selective Fading in Single- and Multi-Carrier Communications

Mr. Cihan Tepedelenlioglu

Dept. of Electrical and Computer Engineering
University of Minnesota
Minneapolis, Minnesota

Thursday, March 29th, 11:00 AM, ENS 637

cihan@ece.umn.edu


Abstract

The tremendous growth of wireless industry along with the limitations of its resources has made modeling and mitigation of the wireless channel a subject of paramount importance. Specifically, joint mitigation of time-selective and frequency-selective fading is an important and challenging problem in mobile communications. Relying on transmitter-induced redundancy using filterbanks, we have developed novel channel estimation and symbol recovery approaches for blind identification and equalization of time- and frequency-selective channels where the time variation is modeled deterministically by a basis expansion. The resulting algorithms enable the usage of a single antenna, dispense with channel disparity conditions of existing approaches, and allow channel order overestimation.

Towards the goal of modeling and mitigation of fading channels, estimation of the statistical channel parameters is also needed. These parameters include, local average power, maximum Doppler spread, the Ricean K factor, shadow variance and correlation distance, all of which are crucial in assessing link quality in wireless systems. We have developed novel estimators for these parameters, and assessed their performance using analytical tools.

The ability of OFDM systems to mitigate frequency-selective channels is also impaired by the presence of a carrier frequency offset (CFO), which is a form of time-selectivity. We have developed a matrix-vector framework for OFDM systems with CFOs in baseband, and used this framework in developing new approaches for CFO estimation that do not suffer from the lack of identifiability problems of current schemes.

Biography

Mr. Cihan Tepedelenlioglu is a graduate student with the Communication and Signal Processing Laboratory at University of Minnesota under Prof. Georgios B. Giannakis. His general interests lie in the area of modeling and equalization of time- and frequency- selective fading channels, and time-varying systems. Recently, he has been investigating how precoding at the input can help in blind identification / equalization problems for fading channels.


A list of Telecommunications and Signal Processing Seminars is available at from the ECE department Web pages under "Seminars". The Web address for the Telecommunications and Signal Processing Seminars is http://anchovy.ece.utexas.edu/seminars