Central limit theorem for quadratic forms based on Gram random matrices

vendredi 4 mars 2011, 13h30 - 14h30

Salle de réunion, espace Turing

Motivated by applications in wireless communications, we are interested by the limiting behavior of the SINR at the output of the LMMSE estimator. The SINR can be written as a quadratic form based on random Gram matrices.
In this work, we prove that, when centered and properly rescaled, this statistic (SINR) satisfies a central limit theorem.