Monday, March 1, 1999
3:00 PM (refreshments 2:45)
Marlar Lounge, Room 37-252
EECS Special Seminar
Abstract
Wireless communications rely on channels which are inherently time-varying and which exhibit intersymbol interference (ISI). Because of time variations, channels cannot be known exactly, but measurements of the channels can be made available to senders and receivers. Information about the channel is commonly termed side information. In this talk, we consider a channel which has side information at both the sender and the receiver. First, we show that, as long as the channel evolution ensures certain decorrelation properties, we can define a maximum reliably achievable data rate (capacity) for such channels. Our approach is novel in that we consider jointly the effects of time-variations and ISI, whereas other approaches have relied on decoupling the two effects. Next, we consider how such capacity can be reached for channels where the channel side information consists of perfect instantaneous knowledge of the channel. Traditional decomposition techniques for ISI channels do not apply. Using our coding theorem, we show that capacity can be achieved and we find the sender's strategy for achieving capacity. Finally, some new research directions are presented.
|
Modified: Feb 26, 1999|
Current events|
Your comments
and inquiries are welcome.