A Distributed Iterative Interference Alignment Scheme for K-user MIMO Interference Channel

In this paper, we propose an iterative interference alignment scheme to improve the sum rate performance of the K-user MIMO interference channel. We design the precoding matrices to minimize the leakage interference. Linear receiver decoding matrices based on the Minimum Mean Square Error (MMSE) criterion are used to further eliminate the interference. Furthermore, the receiver filter is designed not only for removing the interference from the undesired transmitters but also separating its own spatial streams. Our scheme can be achieved with lower complexity comparing with maximizing the signal to interference-plus-noise rate (MAX-SINR) scheme while providing a comparable performance, and it can be done in a distributed manner where only local channel knowledge is needed for each user. By iteratively updating the precoding and decoding matrices, it converges to the solution that can remove the interference of each user effectively. Monte-Carlo simulations have been provided to illustrate the performance of the proposed schemes comparing with the existing schemes.