Limited Feedback Design for Interference Alignment on MIMO Interference Networks with Heterogeneous Path Loss and Spatial Correlations

摘要

Interference alignment is degree of freedom optimal in K -user MIMOinterference channels and many previous works have studied the transceiverdesigns. However, these works predominantly focus on networks with perfectchannel state information at the transmitters and symmetrical interferencetopology. In this paper, we consider a limited feedback system withheterogeneous path loss and spatial correlations, and investigate how thedynamics of the interference topology can be exploited to improve the feedbackefficiency. We propose a novel spatial codebook design, and perform dynamicquantization via bit allocations to adapt to the asymmetry of the interferencetopology. We bound the system throughput under the proposed dynamic scheme interms of the transmit SNR, feedback bits and the interference topologyparameters. It is shown that when the number of feedback bits scales with SNRas C_{s}\cdot\log\textrm{SNR}, the sum degrees of freedom of the network arepreserved. Moreover, the value of scaling coefficient C_{s} can besignificantly reduced in networks with asymmetric interference topology.