Proceedings of IEEE International Symposium on Information Theory (ISIT), June 13th - 18th, 2010.
This paper provides a simple closed-form SINR expression for interference alignment over MIMO channels with channel uncertainty. Assuming linear processing (specifically, zero-forcing) at the transmitters and receivers and a complex Gaussian interference channel, we show that random matrix theory can be successfully applied to find the SINR distribution of each stream for each user with channel uncertainty. Perfect channel knowledge constitutes a special case. This SINR distribution allows easy calculation of useful performance metrics like symbol error-rate and achievable sum rate.[download full paper]