Capacity of Communication Networks with Common Messages
Date and time:
Thursday, October 31, 2013 - 3:30pm
Engineering Center, ECOT 831
To meet the rising demand for high-rate wireless data services, a clear
picture of what communication rates are, or are not, possible would be of
invaluable use. Information theoretic models strive to answer such funda-
mental questions. This talk will focus on simple subsystems (or channels)
of modern communication systems. In particular, we focus on the multiple
access and interference channels. For these channels, we characterize the set
of data rates that can be supported with an arbitrarily low probability of
error (that is, the channel's capacity region). Our models incorporate ele-
ments of modern wireless systems, such as multiple antenna communication
(or MIMO communication), and a simple model of cooperation: the presence
of common messages in addition to private messages.
We develop our results in a general framework that subsumes previous
results into a simpler framework. We provide a novel maximum entropy
principle that may be of use in other information theoretic studies. Finally,
where the characterization of capacity is too terse to provide insight, we
develop an approximate characterization of capacity that is \good enough"
and provides intuition for what performance tradeos exist in the design of
multiaccess communication systems.