Threshold-based selective feedback for opportunistic interference alignment

Opportunistic interference alignment (OIA) exploits channel randomness and multiuser diversity by user selection. In this paper, we address the major disadvantage of OIA which requires the feedback of the locally measured interference alignment from all users. We propose a selective feedback scheme for OIA by thresholding, where only a subset of users are required to send feedback to the transmitter. The proposed scheme can reduce the amount of feedback and still achieve the optimal degrees of freedom (DoF). We characterize the threshold and the corresponding feedback load to achieve the full DoF for a given signal-to-noise ratio. Both theoretical analysis and simulation results show that the amount of feedback can be dramatically reduced (by one order of magnitude at 20dB SNR and two orders of magnitude at 30dB SNR), while still preserving the essential DoF promised by conventional OIA with full feedback.