EIV-Based Interference Alignment Scheme with CSI Uncertainties

Abstract EN

A novel interference alignment (IA) scheme based onthe errors-in-variables (EIV) mathematic model has been proposed toovercome the channel state information (CSI) estimation error forthe MIMO interference channels.

By solving an equivalently unconstrainedoptimization problem, the proposed IA scheme employing a weightedtotal least squares (WTLS) algorithm can obtain the solution to aconstrained optimization problem for transmit precoding (TPC) matricesand minimizes the distortion caused by imperfect CSI according tothe EIV model.

It is shown that the design of TPC matrices can berealized through an efficient iterative algorithm.

The convergenceof the proposed scheme is presented as well.

Simulation results showthat the proposed IA scheme is robust over MIMO interference channelswith imperfect CSI, which yields significantly better sum rate performancethan the existing IA schemes such as distributed iterative IA, maximumsignal-to-interference-plus-noise ratio (Max SINR), and minimum meansquare error (MMSE) schemes.