EECS CSPL SEMINAR SERIES
WINTER TERM 1999


Joseph Kangk

Joseph Kang

EECS Department

University of Michigan

jhkang@eecs.umich.edu



Thursday, March 11
4:30 - 5:30 PM
Room 1001 EECS


Iterative estimation and decoding for channels with memory

Abstract-

Recently, turbo codes have received a lot of notoriety as they have been shown to achieve data communication at signal-to-noise ratios close to the Shannon limit. The excellent performance of turbo codes explains why much of the current research is focused on applying them to different systems.

Channel coding schemes have generally been designed to increase the reliability of information transmission when the errors are statistically independent. However, many channels such as multipath and fading which exhibit bursts of errors. A common method for dealing with these bursts is to interleave the information in such a manner that the channel appears memoryless. Thus if interleaving is applied to a burst channel, a code devised for independent errors can be applied. However, such a scheme does not make use of the information inherent in the memory. Because multiple bits have been transmitted over similar channel conditions, it might be useful to estimate the channel state and use this information in the decoder. In this presentation, we consider a receiver which jointly performs iterative channel estimation and iterative decoding over channels with memory.




return to Previous CSPL Seminars