This article presents limited feedback-based precoder quantization schemes for Interference Alignment (IA) with bounded channel state information (CSI) uncertainty. Initially, this work generalizes the min-max mean squared error (MSE) framework, followed by the development of robust precoder and decoder designs based on worst case MSE minimization. The proposed precoder and decoder designs capture the effect of CSI uncertainty using a single parameter, which is independent of the CSI uncertainty in the direct links. The IA algorithms derived employing these proposed designs are shown to be globally convergent under certain conditions. Moreover, precoder quantization schemes are presented for scenarios with and without CSI uncertainty for practical implementation of these techniques in systems with limited feedback. An optimal bit allocation scheme is presented to maximize the sum rate via analysis of the rate loss upper bound. Simulation results demonstrate the improved performance of the proposed IA schemes for various scenarios considering imperfect CSI as well as limited feedback.
