Achieving Full Sum DoF in the SISO Interference Channel With Feedback Delay

Abstract

Channel state information at the transmitter (CSIT) is of utmost importance to manage interference in multi-user wireless networks, in which transmission rates at a high signal-to-noise ratio are characterized by (sum) degrees of freedom (DoF). Ingenious techniques have been proposed to deal with delayed CSIT, notably a scheme by Maddah-Ali and Tse, providing significant DoF with outdated CSIT. However, with most techniques, any delay in the CSIT feedback still results in a DoF loss. A notable exception is the work by Lee and Heath, allowing one to preserve the optimal DoF in the underdetermined MISO BC and IC, but for feedback delays T fb decreasing with the number of users K, up to a fraction 1/K of the channel coherence time T c . In this work, an ergodic interference alignment scheme that preserves the K/2 sum DoF of the K-user IC is proposed for T fb ≤ T c /2. It is also proven that T fb = T c /2 is the longest for which the optimal K/2 DoF can still be obtained for all K.