Mac Channels

23 Apr 2015

In a joint work with Venkat Anantharam from UC Berkeley, WNCG Prof. François Baccelli  derived the capacity region in the Poltyrev sense of the dimension matched MAC channel. They gave a representation of the error probabilities for each subset of transmitters based on Palm theory, and random coding exponents for each type of error event in the case without power constraints, in the case of independent and identically distributed Gaussian noise, with arbitrary positive definite covariance matrix at each time. They showed that this also leads to random coding error exponents in the traditional power-constrained case, where the power constraint at each transmitter is defined by an arbitrary positive definite matrix at each time. 

This will be presented at IEEE ISIT in Hong Kong in June.