A robust interference alignment algorithm in a MIMO interference channel

The sum rate maximization problem in Interference Alignment (IA) is hard to be solved, for the sum rate function is non-convex. The Weighted Minimum Mean Square Error (WMMSE) algorithm settles this problem by converting the original problem into an equivalent convex one. In this paper, an enhanced version of the WMMSE algorithm is presented, which takes the statistical information of the channel estimation error into consideration. Simulation results show that this algorithm has better performance than the original WMMSE algorithm and another distributed IA algorithm. Furthermore, the impact of time varying error caused by Doppler Shift is also investigated.