An information-estimation perspective of communications over Gaussian interference channels

The information-estimation relationship is used to gain insight into useful coding schemes operating over the Gaussian interference channel. After reviewing basic I-MMSE relations and their implications on point-to-point coding over the Gaussian channel, we focus on the Gaussian interference channel. Here the inflicted interference is measured by the associated minimum mean square error (MMSE). Structure of codes achieving reliable communication at some specific signal-to-noise ratio (snr) and constrained by the permitted MMSE at lower snr values, modeling the interference, are discussed. It is shown that layered superposition codes attain optimal performance, providing thus some engineering insight to the relative efficiency of the Han-Kobayashi coding strategy in the two-user interference scenario. We also address related research challenges.